Whole Blood Transcriptomic Signature of Cushing’s Syndrome

Posted on July 15, 2024 by MaryO

Abstract

Objective

Cushing’s syndrome is characterized by high morbidity and mortality with high interindividual variability. Easily measurable biomarkers, in addition to the hormone assays currently used for diagnosis, could reflect the individual biological impact of glucocorticoids. The aim of this study is to identify such biomarkers through the analysis of whole blood transcriptome.

Design

Whole blood transcriptome was evaluated in 57 samples from patients with overt Cushing’s syndrome, mild Cushing’s syndrome, eucortisolism, and adrenal insufficiency. Samples were randomly split into a training cohort to set up a Cushing’s transcriptomic signature and a validation cohort to assess this signature.

Methods

Total RNA was obtained from whole blood samples and sequenced on a NovaSeq 6000 System (Illumina). Both unsupervised (principal component analysis) and supervised (Limma) methods were used to explore the transcriptome profile. Ridge regression was used to build a Cushing’s transcriptome predictor.

Results

The transcriptomic profile discriminated samples with overt Cushing’s syndrome. Genes mostly associated with overt Cushing’s syndrome were enriched in pathways related to immunity, particularly neutrophil activation. A prediction model of 1500 genes built on the training cohort demonstrated its discriminating value in the validation cohort (accuracy .82) and remained significant in a multivariate model including the neutrophil proportion (P = .002). Expression of FKBP5, a single gene both overexpressed in Cushing’s syndrome and implied in the glucocorticoid receptor signaling, could also predict Cushing’s syndrome (accuracy .76).

Conclusions

Whole blood transcriptome reflects the circulating levels of glucocorticoids. FKBP5 expression could be a nonhormonal marker of Cushing’s syndrome.

Significance

In Cushing’s syndrome, specific hormone assays inform about the level of deviation from normal range. The blood transcriptome signature proposed here is also able to discriminate patients, without any hormone measurements. This direct measurement of the biological impact of glucocorticoids at a tissue level may better reflect the individual consequences of glucocorticoid excess.

Introduction

Cushing’s syndrome (CS) is a condition characterized by chronic cortisol excess related to glucocorticoid treatment (exogenous Cushing’s syndrome) or to endogenous hypercortisolism. The excessive cortisol secretion may be due to either adrenocorticotropic hormone (ACTH)–dependent conditions, most often an ACTH-producing pituitary adenoma (Cushing’s disease), or ACTH-independent causes, commonly a benign adrenal adenoma.1 Chronic exposure to glucocorticoid excess results in specific complications, including cardiovascular and thromboembolic diseases, diabetes mellitus, metabolic syndrome, osteoporosis, and neurocognitive disorders. Numerous comorbidities result in impaired quality of life and increased mortality.2-4

Despite the availability of different hormonal tests for diagnosis and follow-up, the clinical management of these patients remains challenging, since none of the available tools proved to be fully accurate due to the variable pattern of cortisol secretion and the pitfalls of the hormonal immunoassays.5,6 Moreover, the clinical effects of glucocorticoid exposure on peripheral tissues depend not only on the intensity and duration of glucocorticoid excess but also on the peripheral glucocorticoid metabolism and the individual sensitivity to glucocorticoids, not accurately estimated by hormonal parameters. This results in the high interindividual variability frequently reported in Cushing’s syndrome.7,8 Recent studies suggested that the combined assessment of cortisol secretion, cortisone-to-cortisol peripheral activation by the 11β-hydroxysteroid dehydrogenase enzyme, and glucocorticoid receptor sensitizing variants may better estimate the risk to develop each type of complications.9-11

These aspects are crucial mainly for the management of patients with mild Cushing’s syndrome, not clearly characterized by classical features of cortisol excess but consistently associated to an increased risk of morbidities and mortality.12,13 Mild hypercortisolism can occur in different settings. In patients with adrenal incidentalomas, mild hypercortisolism is currently referred to as mild autonomous cortisol secretion (MACS).14 In patients with Cushing’s disease, mild hypercortisolism occurs when hypercortisolism persists/recurs after pituitary surgery or under medical treatment.12,15,16 Irrespective of the origin of cortisol excess, it is still debated whether patients with mild hypercortisolism, as well as those under low-dose systemic or local glucocorticoid therapy, need a close follow-up for cortisol excess–related complications and specific preventive treatments.17-19

In this context, genomic-based studies have recently focused on the identification of blood molecular markers in patients exposed to glucocorticoid excess, aiming to a better individual characterization of these patients. Particularly, DNA methylation profile has been investigated as a potential biological hallmark of glucocorticoid action. Previous studies suggested an association between hypothalamic–pituitary–adrenal axis dysregulation and specific blood DNA methylation profiles, particularly in post-traumatic stress disorders, while recently a dynamic whole blood DNA methylation signature reflecting glucocorticoid excess has been identified.20-22 In both genomic-based and preclinical studies, FKBP5, a gene implicated in glucocorticoid signaling, emerged as potential non hormonal marker of glucocorticoid excess.22-24

The present study completes the previous approaches exploring the impact of glucocorticoids on whole blood transcriptome to better understand the molecular mechanisms of glucocorticoid impregnation. Specifically, through the analysis of whole blood transcriptome profiles from patients with endogenous Cushing’s syndrome, eucortisolism, or adrenal insufficiency, we proposed a transcriptome signature predicting glucocorticoid excess.

Materials and methods

Patients and samples

Fifty-seven blood samples were collected from 43 patients with a confirmed diagnosis of endogenous Cushing’s syndrome, followed in Cochin Hospital (APHP, Paris, France). Diagnostic criteria of Cushing’s syndrome included increased 24-h urinary free cortisol, abnormal cortisol after 1 mg dexamethasone suppression, and altered circadian cortisol rhythm, following consensus guidelines.25

For 14 patients, blood samples were collected before correction of Cushing’s syndrome and at least 3 months after Cushing’s syndrome treatment. At the time of blood sampling, patients were classified as overt Cushing’s syndrome, mild Cushing’s syndrome, eucortisolism, or adrenal insufficiency, depending on clinical and hormonal evaluation. Briefly, overt Cushing’s syndrome patients presented clinical signs and increased 24-h urinary free cortisol (>240 nmol/24 h), increased midnight salivary cortisol (>6 nmol/L), and insufficient cortisol suppression after 1 mg dexamethasone (>50 nmol/L). The mild Cushing’s syndrome cohort included patients with mild hypercortisolism due to either Cushing’s disease or benign adrenal Cushing’s syndrome. The former were patients with persistent or recurrent hypercortisolism after pituitary surgery or during medical treatment; in these patients, the diagnosis of Cushing’s disease was confirmed by the histopathological report consistent with a corticotroph adenoma in the surgically treated patients (6 out of 7) and by the magnetic resonance imaging evidence of a pituitary adenoma in the upfront medically treated patient. Mild hypercortisolism in patients with Cushing’s disease was defined, as previously reported,16,26 by the absence of clinically overt signs of CS and a slight alteration in cortisol secretion, including either increased 24-h urinary free cortisol or increased midnight cortisol or inadequate cortisol suppression after 1 mg of dexamethasone. For mild hypercortisolism due to benign adrenal CS, MACS criteria were used—post-dexamethasone serum cortisol concentration above 50 nmol/L—following recent consensus guidelines.14 The term “mild” was retained for 1 patient with benign adrenal CS who had a borderline dexamethasone suppression test (48 nmol/L) but increased 24-h urinary free cortisol. Eucortisolism was defined as a combination of normalized 24-h urinary free cortisol and of restored cortisol circadian rhythm after either surgery or medical treatment. Adrenal insufficiency was secondary to pituitary surgery for Cushing’s disease. The diagnosis was based on low morning plasma cortisol (<160 nmol/L) and confirmed by the insufficient response to 250 µg corticotropin stimulation test (<500 nmol/L), following the current consensus guidelines.27,28 Detailed hormone values for each sample are provided in Table S1.

Thirty additional samples were collected from patients followed in Hôpital Européen Georges Pompidou Hospital (APHP, Paris, France). These patients presented pheochromocytoma (n = 19) and primary hyperaldosteronism (n = 11; Table S1). The diagnosis was made following the consensus guidelines.29,30

The study was conducted in accordance with the Declaration of Helsinki. Signed informed consent for molecular analysis of blood samples and for access to clinical data was obtained from all patients, and the study was approved by the institutional review board (Comité de Protection de Personnes Ile de France 1, projects 13495 and 13311).

RNA collection and extraction

Whole blood samples were collected into PAXgene Blood RNA Tube (PreAnalytiX, Hombrechtikon, Switzerland), following the manufacturer’s instructions. Total RNA was extracted by using PAXgene Blood RNA Kit, v2 (Qiagen, Hilden, Germany), following the manufacturer’s instructions.

Transcriptome data generation

After RNA extraction, RNA concentrations were obtained using nanodrop or a fluorometric Qubit RNA assay (Life Technologies, Grand Island, NY, USA). The quality of the RNA (RNA integrity number, RIN) was determined on the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) following manufacturer’s instructions.

To construct the libraries, 250 ng of high-quality total RNA sample (RIN > 8) was processed using the Stranded mRNA Prep kit (Illumina, San Diego, CA, USA) according to the manufacturer’s instructions. Briefly, after purification of poly-A–containing mRNA molecules, mRNA molecules were fragmented and reverse-transcribed using random primers. Replacement of dTTP (deoxythymidine triphosphate) by dUTP (deoxyuridine triphosphate) during the second-strand synthesis permitted to achieve the strand specificity. Addition of a single A base to the cDNA was followed by ligation of Illumina adapters. Libraries were quantified on a Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, USA), and profiles were assessed using the DNA High Sensitivity LabChip kit on an Agilent Bioanalyzer (Agilent Technologies). Libraries were sequenced on a NovaSeq 6000 System (Illumina), using 51 base-lengths read in a paired-end mode.

Whole blood methylome data

Among the 57 samples included in the transcriptome analysis, 32 were also used for a methylome analysis recently published.22 For each gene, potentially methylated cytosines-referred to as CpGs- in the promoter regions were defined as CpGs belonging to the TSS1500, TSS200, 5′UTR, and first exon regions. CpG methylation levels were analyzed using M-values generated as previously reported.22

Bioinformatics and statistics

Quality control was performed on raw count matrix, with a target of >5 million reads per sample. All samples passed this control. Illumina adapters were removed using Trimmomatic (v0.39) in paired-end mode.31 Reads were aligned to the reference human genome (GRCh37) and counted using STAR (v2.7.9a).32 Counts were aggregated for transcripts corresponding to the same gene, and only genes with a count sum > 0 in all samples were further considered. Globin genes and sex-related genes were also discarded, as previously published.33

Counts were normalized with DESeq2, using rlog transformation34 (v.1.24.0): raw counts were converted to distributed data structures (dds), and lowly expressed genes were removed using a dds > 1 in at least 3 samples as cutoff, obtaining a final dataset of n = 21 116 and n = 57 samples. The 1500 most variable genes were selected to assess the global data structure by principal component analysis (PCA). Overrepresentation analysis of genes most contributing to PCA components was performed using clusterProfiler package35 (v.3.12.0).

From gene counts, blood cell composition was inferred using the online CIBERSORTx tool (Stanford University 2022),36 with the following parameters: B-mode batch correction, disabled quantile normalization, absolute mode, and n = 500 permutations. For each cell types, a score was generated, reflecting the absolute proportion of each cell type in a mixture.

For supervised differential expression analysis, the edgeR package37 (v.3.26.8) was used to read and preprocess the data before analysis: raw counts were converted to counts per million (CPM), and lowly expressed genes were removed using a CPM > 1 in at least 3 samples as cutoff. To remove heteroscedascity of count data, normalized data were transformed using the voom function.38 Differential expression analysis was performed by applying linear modeling using the limma package39 (v. 3.40.6). Differentially expressed genes were selected using a Benjamin–Hochberg adjusted P < .05 and a logFC > 1 as cutoffs. Overrepresentation analysis of differentially expressed genes was performed using the clusterProfiler package. Of note, the edgeR normalization did not significantly modify the normalized expression levels compared to DESeq2 (gene expression correlation r = .9924, P < 2.2e−16).

For predicting glucocorticoid status from transcriptome, we carried out a Ridge-regularized regression (α = 0) using the 1500 most variable genes, with a 4-fold cross-validation, using the glmnet package40 (v. 4.1-1). The optimization of the 1500 gene predictor was performed on a training cohort of 29 samples, randomly selected from the whole cohort and including 18 samples corresponding to overt Cushing’s syndrome and 11 samples corresponding to either eucortisolism or adrenal insufficiency (patients with mild Cushing’s syndrome were excluded). The accuracy of the 1500 gene predictor was assessed on 2 validation cohorts: a first one (n = 17) including overt Cushing’s syndrome, eucortisolism, and adrenal insufficiency samples, and a second one (n = 30) including pheochromocytoma and primary hyperaldosteronism samples. The latter cohort was used to test the specificity of the predictor, given the different nature of catecholamine excess and primary hyperaldosteronism from Cushing’s syndrome.

Quantitative variable comparisons between groups were performed using Student’s t-test for variables following a normal distribution, or Wilcoxon’s test and Kruskal–Wallis test for variables not following a normal distribution. Quantitative variable correlations were performed using Pearson’s or Spearman’s test according to data distribution. Multivariate logistic regression model including the 1500 gene transcriptome predictor and the neutrophil score was used to test the association with glucocorticoid status. All P-values were 2-sided, and the level of significance was set at .05. All tests were computed in R software environment (3.6.0 version).

Results

Cohort presentation

Fifty-seven blood samples were collected from 43 patients (Table 1;  Table S1). Samples were collected at different time points during the disease, thus reflecting different glucocorticoid status: overt Cushing’s syndrome (n = 28), mild Cushing’s syndrome (n = 11), eucortisolism (n = 10), and adrenal insufficiency (n = 8).
Table 1.

Overall cohort presentation and group comparisons.

Glucocorticoid status Whole cohort
median (IQR)		 Training cohort
median (IQR)		 First validation cohort
median (IQR)		 P-valuea
Samples Total 57 29 17
 Overt Cushing’s syndrome N 28 18 10
Urinary free cortisol
nmol/24 h (<240)		 879.5
(419)		 879.5
(307.5)		 904.5
(5469.25)		 .688
Midnight salivary cortisol
nmol/L (<6)		 14
(12)		 11
(8.5)		 17.5
(27.5)		 .034
Plasma cortisol after 1 mg DST
nmol/L (<50)		 232
(288)		 218
(271)		 232
(460)		 .419
 Mild Cushing’s syndrome N 11 NA NA NA
Urinary free cortisol
nmol/24 h (<240)		 273
(100)		 NA NA NA
Midnight salivary cortisol
nmol/L (<6)		 7
(5.5)		 NA NA NA
Plasma cortisol after 1 mg DST
nmol/L (<50)		 56
(19.75)		 NA NA NA
 Eucortisolism N 10 6 4
Urinary free cortisol
nmol/24 h (<240)		 183
(87.75)		 159
(71.25)		 204
(39.25)		 .521
Midnight salivary cortisol
nmol/L (<6)		 4
(1)		 4
(0)		 4.5
(1.25)		 .797
Plasma cortisol after 1 mg DST nmol/L (<50) 35
(11)		 31
(8.5)		 41.5
(6.5)		 .4
 Adrenal insufficiency N 8 5 3
Early morning plasma cortisol nmol/L (160–500) 95.5
(66.75)		 95.5
(28.25)		 98
(98)		 1
Cortisol after ACTH stimulation nmol/L (<500) 405.5
(165.25)		 435.5
(128.75)		 308
(163)		 .142

Cortisol values are provided as median values with interquartile range (IQR). aWilcoxon’s test comparing training and first validation cohorts.

Median age was 48 years (range: 26 to 73), with a female predominance (2.35 to 1). Cushing’s syndrome corresponded either to Cushing’s disease (n = 26) or to benign adrenal Cushing’s syndrome (n = 17). Mild Cushing’s syndrome cohort included 7 patients with Cushing’s disease and 4 patients with a benign adrenal tumor. Hypercortisolism-related complications, including hypertension, diabetes, and osteoporosis, were present in 41 (71.9%), 16 (28.0%), and 10 (17.5%) patients, respectively.

For the purpose of building and evaluating a glucocorticoid status predictor from blood transcriptome, we focused on patients with overt Cushing’s syndrome, eucortisolism, and adrenal insufficiency, excluding patients with mild Cushing’s syndrome (n = 11) due to their uncertain glucocorticoid status. Patients were randomly assigned either to a training (n = 29) or to a first validation cohort (n = 17). A second validation cohort of 30 samples was used to test the specificity of the predictor, including 19 patients with pheochromocytoma and 11 patients with primary hyperaldosteronism (Table S1).

Impact of glucocorticoid level on whole blood transcriptome

Unsupervised PCA on the 1500 most variable genes of the whole cohort (samples = 57) discriminated patients according to their glucocorticoid status (Figure 1A). This discrimination was mainly based on the first principal component (PC1; Table S2). In terms of gene expression signature, PC1 was enriched in signaling pathways related to immune response, particularly those relative to neutrophils’ activation and degranulation (Figure 1B; Table S3). Beyond the immune response, PC1 was also enriched in genes more generally involved in the response to glucocorticoids,41 including FKBP5, PBX1, SPI1, CDK5R1, CXCL8, NR4A1, and TBX21 (Table S2).

Impact of glucocorticoid levels on whole blood transcriptome. (A) Sample projections based on the combination of the first 2 principal components (PC1 and PC2) of unsupervised PCA performed on the 1500 most variable genes of the whole cohort (n = 57). (B) Dot plot of the 10 most GO-enriched signaling pathways in overt Cushing's syndrome, using the PC1 coefficients.

Figure 1.

Impact of glucocorticoid levels on whole blood transcriptome. (A) Sample projections based on the combination of the first 2 principal components (PC1 and PC2) of unsupervised PCA performed on the 1500 most variable genes of the whole cohort (n = 57). (B) Dot plot of the 10 most GO-enriched signaling pathways in overt Cushing’s syndrome, using the PC1 coefficients.

Accordingly, a supervised comparison of Cushing’s syndrome samples (n = 28) against eucortisolism/adrenal insufficiency samples (n = 18) provided similar results (Figure 2; Table S4).

Differentially expressed genes in overt Cushing's syndrome. Volcano plot of the differentially expressed genes (n = 517) in overt Cushing's syndrome (n = 28) versus eucortisolism/adrenal insufficiency (n = 18).

Figure 2.

Differentially expressed genes in overt Cushing’s syndrome. Volcano plot of the differentially expressed genes (n = 517) in overt Cushing’s syndrome (n = 28) versus eucortisolism/adrenal insufficiency (n = 18).

Predicting glucocorticoid status by blood transcriptome

To predict glucocorticoid status by whole blood transcriptome, we performed a cross-validated Ridge-regularized regression, using the 1500 most variable genes. The 1500 transcriptome predictor was optimized in the training cohort to discriminate overt Cushing’s syndrome from eucortisolism/adrenal insufficiency (Table S5). The predictive value of this model was confirmed on both the first and the second validation cohorts (accuracy of .82 and 1, respectively, Table 2; Table S6). Accordingly, samples from the second validation cohort clustered with eucortisolism/adrenal insufficiency samples, as assessed by PCA (Figure S1).
Table 2.

Performance of molecular predictors, based on the whole blood transcriptome signature and on FKBP5 expression level, in discriminating glucocorticoid excess.

Cohort Predictor Accuracy Sensitivity Specificity
First validation cohort Predictor based on 1500 genes .82 .90 .85
Predictor based on FKBP5 .76 .80 .71
Second validation cohort Predictor based on 1500 genes 1 NAa 1
Predictor based on FKBP5 .46 NAa .46

aNot applicable due to the lack of true positives in the second validation cohort.

Mild Cushing’s syndrome samples—excluded from the training and validation cohorts—were classified either as overt Cushing’s syndrome (n = 5/11, 45.5%) or as eucortisolism/adrenal insufficiency (n = 6/11, 54.5%). Of note, the Ridge scores for samples classified as overt Cushing’s syndrome in the mild Cushing’s syndrome cohort was lower than in the training and the first validation cohorts (Wilcoxon, P = .008). The Ridge scores for samples classified as eucortisolism/adrenal insufficiency in the mild Cushing’s syndrome cohort did not differ from the training and first validation cohorts (Wilcoxon, P = .9; Table S6). Accordingly, mild Cushing’s syndrome samples were projected in-between overt Cushing’s syndrome and eucortisolism samples on PCA (Figure 1A).

We then tested whether the glucocorticoid status could be predicted using a single gene. We focused on FKBP5, due to (1) its Ridge regression coefficient being among the highest (Table S5), (2) its potential ability to discriminate Cushing’s syndrome,22,23 and (3) its known implication in glucocorticoid signaling (Figure 3A).42 The prediction accuracy of FKBP5 expression was comparable to the 1500 gene transcriptome predictor in the first validation cohort (accuracy: .76), but lower in the second validation cohort (accuracy: .46; Table 2; Table S7). The other genes involved in the glucocorticoid response found enriched in PC1 were not further analyzed as potential single biomarkers, since their association with Cushing’s syndrome was not confirmed in supervised analysis, and since their Ridge regression coefficients were lower than FKBP5 coefficient (Table S5).

FKBP5 expression related to the different glucocorticoid status. (A) Boxplot of FKBP5 gene expression in the different study groups. *Student's t-test P < .001. (B) Representation of the positive correlation between the 24-h urinary free cortisol and FKBP5 expression (r = .72, P = 2.032e−10). (C) Representation of the inverse correlation between FKBP5 expression and the mean methylation level (M-value) of FKBP5 promoter–associated CpG site (r = −.86, P = 1.312e−10).

Figure 3.

FKBP5 expression related to the different glucocorticoid status. (A) Boxplot of FKBP5 gene expression in the different study groups. *Student’s t-test P < .001. (B) Representation of the positive correlation between the 24-h urinary free cortisol and FKBP5 expression (r = .72, P = 2.032e−10). (C) Representation of the inverse correlation between FKBP5 expression and the mean methylation level (M-value) of FKBP5 promoter–associated CpG site (r = −.86, P = 1.312e−10).

We then tested the contribution of blood cell composition in the 1500 gene transcriptome predictor. We inferred the different blood cell subtype proportions from the whole blood transcriptome of each sample. An expected increase of neutrophil proportion in overt Cushing’s syndrome43,44 was observed (Kruskal–Wallis’s test, P = 8.5e−06; Table S1 and Figure S2). In a multivariate model combining the 1500 gene transcriptome predictor and the neutrophil score, the 1500 gene transcriptome predictor remained significant (P = .002; Table 3).
Table 3.

Multivariate model combining the 1500 gene transcriptome predictor and neutrophil scores.

Variables OR 95% CI P-value
1500-genes predictor 4.37 2.06–15.3 .002
Neutrophils score .48 .02–6.13 .6

Training and first validation cohorts were combined. Two statuses were considered: overt Cushing’s syndrome and eucortisolism/adrenal insufficiency.

Abbreviations: OR, odds ratio; CI, confidential Interval.

Association between blood transcriptome and Cushing’s syndrome complications

The 1500 gene transcriptome predictor was positively correlated to the 24-h urinary free cortisol (r = .78, P = 2.993e−13; Figure S3). The 1500 gene transcriptome predictor was higher in patients with osteoporosis (Wilcoxon, P = 2.9e−05), while the 24-h urinary free cortisol did not show any difference (Wilcoxon, P-value of .17, Figure 4A and B). No difference was observed between patients with and without diabetes (Wilcoxon, P = .31), nor with or without hypertension (Wilcoxon, P = .25), and the 1500 gene transcriptome predictor was not correlated to body mass index (BMI) (P-value = .108).

Potential markers of osteoporosis in overt Cushing's syndrome. Association between osteoporosis and 24-h urinary free cortisol (A), 1500 gene transcriptome predictor (B), and FKBP5 expression (C). For 24-h urinary free cortisol, values are expressed as log10.

Figure 4.

Potential markers of osteoporosis in overt Cushing’s syndrome. Association between osteoporosis and 24-h urinary free cortisol (A), 1500 gene transcriptome predictor (B), and FKBP5 expression (C). For 24-h urinary free cortisol, values are expressed as log10.

Similar findings were obtained with FKBP5 expression level, including a positive correlation with the 24-h urinary free cortisol (r = .72, P = 2.032e−10, Figure 3B), a higher expression in patients with osteoporosis (Wilcoxon, P = 2.9e−05; Figure 4C), no difference in patients with diabetes (Wilcoxon, P = .72) or hypertension (Wilcoxon, P = .4), and no correlation with BMI (P = .657).

Association of whole blood transcriptome with whole blood methylome

For 32 samples with both whole blood transcriptome and methylome22 available (n = 32), a correlation analysis was performed. A majority of genes differentially expressed in overt Cushing’s syndrome showed a negative correlation with CpG sites of their promoter regions (Table S8). FKBP5 was among the genes showing the strongest inverse correlation (r = − .86, P adjusted = 5.94e−09; Figure 3C).

Discussion

In this study, we identified a whole blood transcriptome signature predicting the glucocorticoid excess. This signature, in addition to the hormone assays currently used for diagnosis, could reflect the individual biological impact of glucocorticoids.

We designed a predictor with optimal selection of transcriptome biomarkers able to differentiate overt Cushing’s syndrome from eucortisolism and adrenal insufficiency. The predictive value of such transcriptome predictor was confirmed on 2 validation cohorts. For patients with mild Cushing’s syndrome, our predictor showed intermediate classification, confirming the clinical heterogeneity of this group. Indeed, these intermediate patients indisputably fall in-between patients with overt Cushing’s syndrome and eucortisolism, with some overlap in both groups. Whether such non hormonal biomarkers, directly measuring glucocorticoid action, can be useful for the specific management of these patients remains to be established. The question is important, considering the high prevalence of mild Cushing’s syndrome in the general population and the still-ongoing debate on complications’ surveillance and treatment of choice.45 Here, a proper evaluation of mild Cushing’s syndrome is difficult, due to both the lack of a clear clinical definition and to the size of the cohort, not large enough to assess the existence of a specific signature for these patients, thus representing a limitation of this study. Another open question is whether the markers presented here would have comparable relevance in patients with exogenous Cushing’s syndrome, related to glucocorticoid treatments, especially for the common situation of long-term treatment with low glucocorticoid doses or with “local” glucocorticoid treatments.

Noteworthy, this identified signature derives from whole blood, a mixture of various cell types with potentially cell-dependent impact of glucocorticoids on transcriptome profile. Indeed, glucocorticoids have a direct effect on white blood cell count inducing an increase in the neutrophil proportion.43,44 We inferred white blood cell count from transcriptome profile for each sample, and, as expected, overt Cushing’s syndrome samples were characterized by higher neutrophil score, and, accordingly, genes differentially expressed in this group were enriched in immunity-related pathways, mainly in the activation and degranulation of neutrophils. However, among the genes differentially expressed in overt Cushing’s syndrome, we also identified genes more specifically involved in glucocorticoid response, suggesting differences not only related to immunity. Moreover, we demonstrated that the prediction based on transcriptome signature remained significant after adjustment for neutrophil score and therefore that transcriptome profile does not only reflect blood composition variations.

Whole blood transcriptome analysis is not easily reproducible in clinical practice. Thus, we tried to simplify the marker by focusing on one single gene. FKBP5, as a potential surrogate of the 1500 gene transcriptome signature, was able to differentiate and predict Cushing’s syndrome with a good accuracy. FKBP5 (FK506-binding protein 51) is a co-chaperone of heat shock protein 90 (Hsp90) involved in the regulation of the glucocorticoid receptor activity, maintaining it unbound and inactive in the cytoplasm, thus restricting the nuclear translocation of the cortisol receptor complex.24,46 According to preclinical studies, in the presence of glucocorticoid excess, FKBP5 expression increases at both mRNA and protein levels as an effect of intracellular negative feedback.47 Previous studies also showed that FKBP5 expression is sensitive to exogenous glucocorticoids in healthy volunteers and that FKBP5 levels are higher in patients with Cushing’s syndrome, while decreasing to normal baseline levels after successful surgery.23 It has been also demonstrated that the methylation of FKBP5 is affected by stress and dynamically by glucocorticoid level in patients with endogenous Cushing’s syndrome.42 Of note, in our second validation cohort, including patients with pheochromocytoma and primary aldosteronism, the ability of FKBP5 expression level to properly call the absence of Cushing’s syndrome dropped compared to the first validation cohort, raising concerns about potential limits in specificity. These results also highlight the importance of using larger validation cohorts with a wide variety of conditions before using such a biomarker in routine.

Interestingly, in patients with overt Cushing’s syndrome, beyond the correlation between gene expression and 24-h urinary free cortisol, the variability of gene expression was higher in patients with moderate increase of 24-h urinary free cortisol. This suggests a potential informative role of gene expression markers in patients with moderate cortisol increase. In this line, Guarnotta et al. showed that the level of urinary hypercortisolism does not seem to correlate with Cushing’s syndrome severity and that clinical features and cortisol excess–related comorbidities are more reliable indicators in the assessment of disease severity.48 In our study, the transcriptomic profile could discriminate Cushing’s syndrome patients with and without osteoporosis, although the 24-h urinary free cortisol values did not differ between the two groups. However, these results need additional validation, due to the limited cohort size and because of potential confounders not considered, including pre-existing diagnosis of osteoporosis and other determinants of skeletal fragility. Although this preliminary finding further supports the potential value of gene expression markers in predicting catabolic complications, to which extent these biomarkers are relevant in clinical practice remains to be established and better explored in larger cohorts of patients with moderate Cushing’s syndrome.

The transcriptome profile identified in this study also confirmed the previous findings obtained by analyzing the whole blood methylome in Cushing’s syndrome. The negative correlation between promoter methylation and gene expression strengthens our results and underlines the importance of epigenetic alterations in Cushing’s syndrome.49

In conclusion, we showed that the whole blood transcriptome reflects the circulating levels of glucocorticoids and that FKBP5 expression level could be a single gene non hormonal marker of Cushing’s syndrome.

Acknowledgments

We thank the Genomic platform and the team “Genomic and Signaling of Endocrine Tumors” of Institut Cochin, the French COMETE research network, the European Network for the Study of Adrenal Tumor (ENSAT), and the European Reference Network on Rare Endocrine Conditions (Endo-ERN).

Supplementary material

Supplementary material is available at European Journal of Endocrinology online.

Funding

This project has received funding from the European Union’s Horizon 2020 Research and Innovation program under grant agreement no. 633983 and the Programme Hospitalier de Recherche Clinique “CompliCushing” (PHRC AOM 12-002-0064). This work was also supported by the Programme de Recherche Translationnelle en Cancérologie to the COMETE network (PRT-K COMETE-TACTIC).

Authors’ contribution

Maria Francesca Birtolo (Data curation [equal], Formal analysis [equal], Writing—original draft [equal]), Roberta Armignacco (Conceptualization [equal], Data curation [equal], Formal analysis [equal], Writing—review & editing [equal]), Nesrine Benanteur (Formal analysis [equal]), Bertrand Baussart (Writing—review & editing [equal]), Chiara Villa (Writing—review & editing [equal]), Daniel De Murat (Formal analysis [equal]), Laurence Guignat (Writing—review & editing [equal]), Lionel Groussin (Writing—review & editing [equal]), Rosella Libé (Writing—review & editing [equal]), Maria-Christina Zennaro (Data curation [equal], Writing—review & editing [equal]), Meriama Saidi (Data curation [equal]), Karine Perlemoine (Data curation [equal]), Franck Letourneur (Data curation [equal]), Laurence Amar (Data curation [equal], Writing—review & editing [equal]), Jérôme Bertherat (Writing—review & editing [equal]), Anne Jouinot (Conceptualization [equal], Formal analysis [equal], Writing—original draft [equal]), and Guillaume Assié (Conceptualization [equal], Formal analysis [equal], Funding acquisition [equal], Project administration [equal], Writing—original draft [equal]).

Data availability

Transcriptome data generated and analyzed in this study are available in the EMBL-EBI BioStudies repository (reference number: S-BSST1241).

Author notes

Conflict of interest: G.A. is on the editorial board of EJE. G.A. was not involved in the review or editorial process for this paper, on which he is listed as an author.

© The Author(s) 2024. Published by Oxford University Press on behalf of European Society of Endocrinology.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

 

Filed under: adrenal crisis, Cushing's, Diagnostic Testing | Tagged: , , , , , , | Leave a comment »

Epigenetic Mechanisms Modulated by Glucocorticoids With a Focus on Cushing Syndrome

Posted on April 19, 2024 by MaryO

Abstract

In Cushing syndrome (CS), prolonged exposure to high cortisol levels results in a wide range of devastating effects causing multisystem morbidity. Despite the efficacy of treatment leading to disease remission and clinical improvement, hypercortisolism-induced complications may persist. Since glucocorticoids use the epigenetic machinery as a mechanism of action to modulate gene expression, the persistence of some comorbidities may be mediated by hypercortisolism-induced long-lasting epigenetic changes. Additionally, glucocorticoids influence microRNA expression, which is an important epigenetic regulator as it modulates gene expression without changing the DNA sequence. Evidence suggests that chronically elevated glucocorticoid levels may induce aberrant microRNA expression which may impact several cellular processes resulting in cardiometabolic disorders.

The present article reviews the evidence on epigenetic changes induced by (long-term) glucocorticoid exposure. Key aspects of some glucocorticoid-target genes and their implications in the context of CS are described. Lastly, the effects of epigenetic drugs influencing glucocorticoid effects are discussed for their ability to be potentially used as adjunctive therapy in CS.

epigenetic, glucocorticoids, Cushing syndrome
Issue Section:

Mini-review

In Cushing syndrome (CS), adrenocorticotropic hormone (ACTH) hypersecretion by a pituitary adenoma or an ectopic source, or autonomous cortisol hypersecretion by an adrenal tumor, induces chronic endogenous hypercortisolism with loss of the cortisol circadian rhythm (1). CS is more prevalent in women than men and frequently occurs in the fourth to sixth decades of life (2).

Glucocorticoids (GC) have extensive physiological actions and regulate up to 20% of the expressed genome, mainly related to the immune system, metabolic homeostasis, and cognition. Therefore, the prolonged exposure to high cortisol levels results in a wide range of devastating effects, including major changes in body composition (obesity, muscle atrophy, osteoporosis), neuropsychiatric disturbances (impaired cognition, depression, sleep disturbances), the metabolic syndrome (obesity, hypertension, insulin resistance, and dyslipidemia), hypercoagulability, and immune suppression (3, 4). The consequences of hypercortisolism lead to compromised quality of life and increased mortality rate (5). The mortality rate in patients with CS is 4 times higher than the healthy control population (6). Risk factors such as obesity, diabetes, and hypertension contribute to the increased risk of myocardial infarction, stroke, and cardiac insufficiency. As a result, cardiovascular disease is the leading cause of the premature death in CS (5). Infectious disease is also an important cause of death in CS (5). Therefore, prompt treatment to control hypercortisolism is imperative to prevent complications and an increased mortality rate.

Despite the efficacy of treatment leading to disease remission, the clinical burden of CS improves, but does not completely revert, in every patient (7). Indeed, obesity, neuropsychiatric disturbances, hypertension, diabetes, and osteoporosis persist in a substantial number of biochemically cured patients. For instance, in a study involving 118 CS patients in remission for about 7.8 years (median), resolution of comorbidities such as diabetes occurred in only 36% of cases, hypertension in 23% of cases, and depression in 52% of the cases (8). It has been proposed that epigenetic changes as a consequence of hypercortisolism is a mechanism of the persistence of some comorbidities (9-12).

Epigenetics is a reversible process that modifies gene expression without any alterations in DNA sequence; frequently it is mediated by histone modification and DNA methylation together with microRNAs (13-15). GCs use the epigenetic machinery as a mechanism of action to regulate gene expression in physiological circumstances, such as metabolic actions and stress response. Its networks involve DNA and histone modifying enzymes, such as DNA methyltransferases (DNMTs), histone acetyltransferases (HATs), and histone deacetylases (HDACs) (16). (Fig. 1) The DNA methylation process catalyzed by DNMTs is usually associated with downregulation of gene expression (17). Histone modifications catalyzed by HAT enzymes induce gene transcription, while those by HDAC enzymes induce transcriptional repression (17). Drugs interfering with these enzymes (so-called epigenetic drugs) may affect the GC genomic actions confirming the interaction between GC and the epigenetic system (18, 19). Furthermore, GC can modulate HDAC and DNMT expression and activity (16, 19, 20). Based on these data it might be speculated that in CS, epigenetic modifications induced by long-term GC exposure plays a role in the development of the disease-specific morbidity (9, 10).

Figure 1.
Glucocorticoid (GC) and its epigenetic machinery. GC through its receptor interacts with DNA and histone modifying enzymes, such as DNA methyltransferases (DNMTs), histone acetyl transferases (HATs), and histone deacetylases (HDAC) to modulate gene expression.

Open in new tabDownload slide

Glucocorticoid (GC) and its epigenetic machinery. GC through its receptor interacts with DNA and histone modifying enzymes, such as DNA methyltransferases (DNMTs), histone acetyl transferases (HATs), and histone deacetylases (HDAC) to modulate gene expression.

In this review we provide an overview of epigenetic aspects of GC action in physiological conditions and in the context of CS. We start with a detailed characterization of how GC, using the epigenetic system, can change chromatin structure in order to activate or silence gene expression. (Fig. 2) Subsequently, we describe the role of epigenetic mechanisms in the regulation of expression of several GC-target genes related to CS. Finally, we present the current evidence of epigenetic changes caused by the long-term of GC exposure and the potential use of epidrugs influencing GC actions.

Figure 2.
Epigenetic mechanisms of the glucocorticoid action to regulate gene expression. The GR is located in cytoplasm in a multi-protein complex; after GC binding, GR dissociates from the multi-protein complex, crosses the nuclear membrane, dimerizes, and binds to the GRE of the target gene. One of the mechanisms of action of GC is through the recruitment of co-regulators together with epigenetic enzymes, such as HAT, to change the chromatin structure, resulting in activation of gene transcription. Also, GR decreases gene expression by tethering other transcriptional factors and recruiting HDAC2, causing histone deacetylation, which leads to a repressed chromatin. GC can cause hypomethylation through downregulation in the expression of DNMT1. Abbreviations: Ac, acetylation; DNMT1, DNA methyltransferase 1; GC, glucocorticoid; GR, glucocorticoid receptor; GRE, glucocorticoid responsive elements; HAT, histone acetyltransferase; HDAC, histone deacetylases; Me: methylation.

Open in new tabDownload slide

Epigenetic mechanisms of the glucocorticoid action to regulate gene expression. The GR is located in cytoplasm in a multi-protein complex; after GC binding, GR dissociates from the multi-protein complex, crosses the nuclear membrane, dimerizes, and binds to the GRE of the target gene. One of the mechanisms of action of GC is through the recruitment of co-regulators together with epigenetic enzymes, such as HAT, to change the chromatin structure, resulting in activation of gene transcription. Also, GR decreases gene expression by tethering other transcriptional factors and recruiting HDAC2, causing histone deacetylation, which leads to a repressed chromatin. GC can cause hypomethylation through downregulation in the expression of DNMT1. Abbreviations: Ac, acetylation; DNMT1, DNA methyltransferase 1; GC, glucocorticoid; GR, glucocorticoid receptor; GRE, glucocorticoid responsive elements; HAT, histone acetyltransferase; HDAC, histone deacetylases; Me: methylation.

Search Strategy

A search of the PubMed database was conducted using the advanced search builder tool for articles in the English language on the following terms “glucocorticoids,” “glucocorticoid receptor,” “Cushing,” “hypercortisolism,” “epigenetic,” “DNA methylation,” “histone deacetylase,” “histone acetyltransferase,” “microRNA” “fkbp5,” “clock genes,” and “POMC.” Moreover, references were identified directly from the articles included in this manuscript. The articles were selected by the authors after being carefully analyzed regarding their importance and impact.

Epigenetic Aspects of Genomic Action of Glucocorticoids

GCs regulate gene expression positively or negatively. GC-responsive genes include genes encoding for proteins associated with inflammation, metabolic processes, blood pressure and fluid homeostasis, apoptosis, cell cycle progression, circadian rhythm, and intracellular signaling (21).

The GC actions are cell type–specific (22). For instance, in an in vitro study, the comparison of GC-expressed genes between 2 cell lines, corticotroph (AtT20) and mammary (3134) cell lines, showed a different set of GC-regulated genes, revealing the cell type–specific nature of GC effects (23). GC function depends on the accessibility of glucocorticoid receptor (GR)-binding sites in the DNA of the target tissue, which in turn is mostly established during cell differentiation. Therefore, different chromatin organization explains the distinct GR-binding sites among different tissues (22, 24, 25). The chromatin accessibility is determined by histone modifications such as acetylation, methylation, phosphorylation, and/or DNA methylation, processes that are both dynamic and reversible (26).

Furthermore, gene expression is regulated in a GC-concentration-dependent manner which is tissue-specific. Only a few genes can be upregulated or downregulated at low concentrations of GC. For example, a dose of dexamethasone (Dex) as low as 0.5 nM selectively activated PER1 (period 1, transcription factor related to circadian rhythm) expression in lung cancer (A549) cells (21, 27). Additionally, continuous GC exposure or pulsed GC (cortisol fluctuation during circadian rhythm) may cause different responses with respect to gene expression (26, 28). For example, constant treatment with corticosterone induced higher levels of PER1 clock gene mRNA expression compared with pulsatile treatment, as demonstrated in an in vitro study using 3134 cell line (28).

The time course for gene expression in response to Dex is fast, with repression occurring slightly slower compared to activation. Half of activated and repressed genes are detected within, respectively, about 40 minutes and 53 minutes following Dex exposure (21).

In short, the transcriptional output in response to GC depends on cell type, as well as on the duration and intensity of GC exposure (21, 24, 26, 27). GCs act as a transcriptional regulatory factor resulting in activating or repressing the expression of genes. The GC exerts its function through binding to corticosteroid receptors, specifically, the mineralocorticoid receptor and the GR, members of the nuclear receptor superfamily (29, 30).

Glucocorticoid Receptor

The GR is located in the cytoplasm in a chaperone complex which includes heat-shock proteins (70 and 90) and immunophilins (such as FK506 binding protein [FKBP5]). Cortisol diffuses across the cell membrane and binds with high affinity to the GR. The activated GR bound to GC dissociates of the multi-protein complex and is transferred to the nucleus, where it ultimately regulates gene expression (26, 31).

GR is a transcription factor encoded by nuclear receptor subfamily 3, group C member 1 (NR3C1) gene, located in chromosome 5, and consisting of 9 exons. It is composed of 3 major functional domains, namely a DNA binding domain (DBD), the C-terminal ligand-binding domain (LBD) and the N-terminal domain (NTB). The LBD recognizes and joins the GC. NTB contains an activation function-1 (AF1) which connects with co-regulators and the members of the general transcription machinery to activate target genes. The DBD comprises 2 zinc fingers motifs that are able to identify and bind to glucocorticoid responsive elements (GREs) (32, 33).

GRα is the most expressed and functionally active GR. GRβ is another isoform which is the result of an alternative splicing in exon 9 of the GR transcript. The difference between the 2 isoforms is the distinct ligand-binding domain in GRβ. This variance prevents the GRβ from binding to GC. In fact, the GRβ counteracts GRα function by interfering with its binding to a GRE in the target gene, and GRβ expression is associated with GC resistance (32). In addition, GRβ has its own transcriptional activity which is independent and distinct from GRα (34).

Another splice variant of human GR, GRγ, is associated with GC resistance in lung cell carcinoma and childhood acute lymphoblastic leukemia (33, 35). There is an additional amino acid (arginine) in the DBD of the GRγ that reduces, by about half, the capacity to activate or suppress the transcription of the target gene, as compared with GRα (32). One study identified GRγ in a small series of corticotroph adenomas (36).

Glucocorticoid Mechanism of Action

The GR-GC complex induces or represses gene expression directly by binding to DNA, indirectly by tethering other transcription factors or yet in a composite manner that consists in binding DNA in association with binding to other co-regulators (35, 37).

The GR has the ability to reorganize the chromatin structure to become more or less accessible to the transcriptional machinery. In the classical mechanism of direct induction of gene expression, the GR dimerizes and binds to a GRE in DNA. The receptor recruits co-regulators, such as CREB binding protein, which has intrinsic histone acetyltransferase (HAT) activity that modifies the chromatin structure from an inactive to an active state. This model, called transactivation, upregulates the expression of some genes related to glucose, protein, and fat metabolism. Gene repression, on the other hand, is accomplished by GR binding to a negative GRE (nGRE) leading to the formation of a chromatin remodeling complex composed by co-repressor factors, such as NCOR1 and SMRT, and histone deacetylases (HDACs), that ultimately turn chromatin less accessible and suppress gene transcription. The gene repression through direct binding events occurs less frequently when compared to gene induction (25, 35, 38).

Another mechanism of GC action is through binding to other transcription factors (tethering). In case of switching off inflammatory genes, GR binds to transcriptional co-activator molecules, such as CREB binding protein with intrinsic HAT activity, and subsequently recruits HDAC2 to reverse histone acetylation, thus resulting in a suppression of the activated inflammatory gene (39). In the same model, GC interacts with other cofactors, such as the STAT family, to induce chromatin modifications resulting in increased gene expression (26).

Furthermore, the transcriptional dynamics of some genes follow a composite manner. In this model, GR, in conjunction with binding to GRE, also interacts with cofactors in order to enhance or reduce gene expression (35).

GCs can also modulate gene expression by influencing the transcription of epigenetic modifiers. An experimental study demonstrated that GC mediated the upregulation of HDAC2 in rats exposed to chronic stress, which in turn decreased the transcription of histone methyltransferase (Ehmt2) that ultimately upregulated the expression of Nedd4. Nedd4 is a ubiquitin ligase, expression of which has been related to cognitive impairment (40). Additionally, GC was found to interact with another epigenetic eraser, namely JMJD3, a histone demethylase, suppressing its transcription in endothelial cells treated with TNFα that led to decreased expression of other genes related to the blood-brain barrier (41).

GCs have the ability to induce (de)methylation changes in DNA, ultimately affecting gene expression. The DNA methylation process triggered by GC involves the family of DNA methyltransferases (DNMT) and ten-eleven translocation (TET) protein (20, 42-44). The DNMT, DNMT1, DNMT3A, and DNMT3B are able to transfer a methyl group to a cytosine residue in DNA, forming 5-methylcytosine (5mC), which negatively impacts gene expression. In contrast, TET protein chemically modifies the 5mC to form 5-hydroxymethylcytosine (5hmC), which ultimately leads to unmethylated cytosine, positively influencing gene expression (45).

Glucocorticoids mainly induce loss of methylation events rather than gain of methylation across the genome (11, 46). The DNA demethylation process can be either active or passive. The active mechanism is linked to the upregulation of TET enzyme expression that follows GC treatment, which was described in retinal and osteocyte cell line model studies (42, 43). The passive demethylation event involves the downregulation (Fig. 2) or dysfunction of DNMT1. DNMT1 is responsible for maintaining the methylation process in dividing cells (45). In case of GC exposure, GC can cause hypomethylation through downregulation in the expression of DNMT1, a process described in the AtT20 corticotroph tumor cell model, or through GC hindering DNMT activity, particularly DNMT1, as demonstrated in the retinal cell (RPE) line (20, 42, 44).

Glucocorticoid-Induced Epigenetic Changes

There are several molecular mechanisms connecting GR activation and epigenetic modifications ultimately affecting gene expression (Fig. 2). As described above, GC uses epigenetic machinery, such as DNA and histone modifying enzymes, to restructure the chromatin in order to induce or silence gene transcription (16, 47).

In an in vitro study using murine AtT20 corticotroph tumor and neuronal cell lines, after chronic GC exposure followed by a recovery period in the absence of GC, the cells retained an “epigenetic memory” with persistence of loss of methylation content in FKBP5 gene but with no increased gene expression at baseline. The functionality of this “epigenetic memory” only became evident in a second exposure to GC, when the cells responded sharply with a more robust expression of FKBP5 gene compared to the cells without previous exposure to GC (44). Another in vitro study, using a human fetal hippocampal cell line, confirmed long-lasting DNA methylation changes induced by GC. The cells were treated for 10 days with dexamethasone, during the proliferative and cell differentiation phases of the cell line, followed by 20 days without any treatment. The second exposure to GC resulted in an enhanced gene expression of a subset of GC-target genes (48). Additionally, using an animal model subjected to chronic stress, a distinct gene expression profile was demonstrated in response to acute GC challenge compared to those without chronic stress history. The proposed mechanism was that chronic stress resulted in GC-induced enduring epigenetic changes in target genes, altering the responsiveness to a subsequent GC exposure (49).

In general, it seems that the majority of differential methylation regions (DMRs) induced by GC are loss of methylation rather than gain of methylation. In an experimental study, an association between hypomethylation and GC exposure was demonstrated in mice previously exposed to high levels of GC. Further analysis demonstrated that the genes linked with DMR were mostly related to metabolism, the immune system, and neurodevelopment (11).

Human studies have also shown that excess of cortisol can induce modifications in DNA methylation. DNA methylation data obtained from whole blood samples from patients with chronic obstructive pulmonary disease (COPD) treated with GC revealed DMR at specific CpG dinucleotides across the genome. These DMR were confirmed by pyrosequencing and annotated to genes, such as SCNN1A, encoding the α subunit of the epithelial sodium channel, GPR97, encoding G protein coupled receptor 97, and LRP3, encoding low-density lipoprotein receptor-related protein 3 (50). Furthermore, it has been proposed that the negative impact of chronic GC exposure on the immune system, which increases the risk of opportunistically infections, may be epigenetically mediated (51). In a clinical study, using whole blood samples, an analysis of genome-wide DNA methylation was performed on patients before and after exposure to GC (51). Long-term GC exposure disrupts, through a persistent modification of the cytosine methylation pattern, the mTORC1 pathway which affects CD4+ T cell biology (51).

Taken together, these data clearly show the interplay between GC signaling and methylation and histone modifications processes suggesting that GC interferes in the epigenetic landscape modulating gene expression. It is possible that most of these GC-induced epigenetic events are dynamic and temporary, while others may persist leading to long-lasting disorders. Further research to provide insight into what makes some events reversible is warranted.

Epigenetic Changes as a Consequence of Long-Term Glucocorticoid Exposure in Cushing Syndrome

The comorbidities associated with CS are associated with increased mortality mainly due to cardiovascular events (52). GC-induced comorbidities in CS may be at least in part epigenetically mediated. Previous study using whole blood methylation profile demonstrated that specific hypomethylated CpG sites induced by GC were associated with Cushing comorbidities, such as hypertension and osteoporosis (46). The study identified a methylator predictor of GC excess which could be used as a biomarker to monitor GC status (46).

The long-term exposure to high cortisol levels may be crucial for the persistence of some morbidities in CS through epigenetic changes. Hypercortisolism-induced persistent changes in visceral adipose tissue gene expression through epigenetic modifications was investigated in a translational study (12). This study combined data from patients with active CS and data from an animal model of CS in active and remitted phase. Interestingly, the study demonstrated long-lasting changes in the transcriptome of adipose tissue that were associated with histone modifications induced by GC. Therefore, these epigenetic fingerprints observed even after the resolution of hypercortisolism may elucidate the mechanism of persistent modifications in gene expression in the visceral adipose tissue (12).

With regard to the persistence of GC-induced DMR, a genome-wide DNA methylation analysis showed a lower average of DNA methylation in patients in remission of CS compared to controls. Interestingly, the most common biologically relevant affected genes were retinoic acid receptors, thyroid hormone receptors, or hormone/nuclear receptors, important genes related to intracellular pathways and regulators of gene expression (9).

In summary, this large body of evidence supports the concept that prolonged GC exposure modulates the epigenetic landscape across the genome by inducing DMR and histone modifications. Some epigenetic modifications are persistent, and this may partially explain the incomplete reversibility of some of CS features following clinical remission.

Glucocorticoid-Target Genes in Cushing Syndrome

A detailed identification and characterization of GC-target genes may shed light in the understanding of the pathophysiology and treatment response in patients with CS. For instance, the GC regulation of pro-opiomelanocortin (POMC) expression as part of the physiologic GC negative feedback may be impaired in Cushing disease (CD), which is an important mechanism for the maintenance of high GC levels (53). Another example is the interaction between GC and clock genes, which may interfere in the loss of the GC circadian rhythm and may contribute to metabolic disorders in CS (54). Furthermore, the suppressive action of GC on drug targets, such as the somatostatin receptor (subtype 2), may influence the efficacy of first-generation somatostatin receptor ligands in normalizing cortisol levels in CD (55). Here we describe how GCs using epigenetic machinery influence the expression of important target genes and their implications in CS.

FKBP5

FK506 binding protein (FKBP5) plays an important role in the regulation of hypothalamic-pituitary-adrenal (HPA) system (56). As part of the GC negative feedback loop, GC binds to hypothalamic and pituitary GR. In the cytoplasm, GR is bound to a multi-protein complex including FKBP5. FKBP5 modulates GR action by decreasing GR binding affinity to GC and by preventing GR translocation from cytoplasm to nucleus (57, 58). In other words, an increase of FKBP5 expression is inversely correlated with GR activity and results in GC resistance leading to an impaired negative feedback regulation in the HPA axis (59).

FKBP5 is a GC-responsive gene; its upregulation by GC is part of an intracellular negative short-feedback loop (60). The mechanism by which GC regulates FKBP5 expression was shown to include inhibition of DNA methylation (44). In a model for CS, mice treated with corticosterone for 4 weeks had a reduced level of DNA methylation of FKBP5 in DNA extracted from whole blood, which was strongly correlated in a negative manner with GC concentration. Interestingly, a negative correlation was also observed between the degree of FKBP5 gene methylation measured at 4 weeks of GC exposure and the percentage of mice visceral fat (61). Accordingly, previous studies have provided compelling evidence of decreased methylation in the FKBP5 gene in patients with active CS compared to healthy control (10, 46). Even in patients with CS in remission, previous data have suggested a small decrease in FKBP5 methylation levels compared to healthy controls (9, 10). In an in vitro study, it was demonstrated that, by decreasing DNMT1 expression, GC is able to reduce FKBP5 methylation levels and, therefore, increase its expression (44).

Likewise, FKBP5 mRNA is also sensitive to GC exposure. A time-dependent increase in blood FKBP5 mRNA after single-dose prednisone administration has been demonstrated in healthy humans (62). Accordingly, patients with ACTH-dependent CS had higher blood FKBP5 mRNA levels compared with healthy controls, and after a successful surgery, FKBP5 mRNA returned to baseline levels (63). Furthermore, in another study, blood FKBP5 mRNA was inversely correlated with FKBP5 promoter methylation and positively correlated with 24-hour urine free cortisol (UFC) levels in patients with CS (46). Taken together, this fine-tuning of FKBP5 DNA methylation and mRNA according to the level of GC suggests that FKBP5 can be used as a biomarker to infer the magnitude of GC exposure.

POMC and Corticotropin-Releasing Hormone

The partial resistance of the corticotroph adenoma to GC negative feedback is a hallmark of CD. Indeed, the lack of this inhibitory effect constitutes a method to diagnose CD, that is, with the dexamethasone suppression test. One of the mechanisms related to the insensitivity to GC can be attributed to GR mutations which are, however, rarely found in corticotrophinomas (64). Another mechanism that was uncovered in corticotroph adenomas is an overexpression of the HSP90 chaperone resulting in reduced affinity of GR to its ligand and consequently GR resistance (53, 65).

In addition, the loss of protein expression of either Brg1, ATPase component of the SWI/SNF chromatin remodeling complex, or HDAC2 has been linked to GC resistance in about 50% of some adenomas (66). The trans-repression process on POMC transcription achieved by GC involves both the histone deacetylation enzyme and Brg1. One mechanism of corticotropin-releasing hormone (CRH)-induced POMC expression is through an orphan nuclear receptor (NR) related to NGFI-B (Nur77). NGFI-B binds to the NurRE sequence in the promoter region of POMC gene and recruits a co-activator to mediate its transcription. In a tethering mechanism, the GR directly interacts with NGFI-B to form a trans-repression complex, which contains the GR itself, Brg1, the nuclear receptor, and HDAC2; the latter being essential to block the gene expression through chromatin remodeling process (53, 66).

In CD, hypercortisolism exerts a negative feedback at CRH secretion from the hypothalamus (67). The mechanism involved in GR-induced suppression of CRH expression is through direct binding to a nGRE in the promoter region of CRH gene and subsequent recruitment of repressor complexes. In a rat hypothalamic cell line, it was demonstrated that Dex-induced CRH repression occurs through coordinated actions of corepressors involving Methyl-CpG-binding protein 2 (MeCP2), HDAC1, and DNA methyltransferase 3B (DNMT3B). Possibly, GR bound to nGRE recruits DNMT3B to the promoter in order to methylate a specific region, subsequently binding MeCP2 on these methylated sites followed by the recruitment of chromatin modify corepressor HDAC1, ultimately resulting in CRH suppression. Another possibility is that 2 independent complexes, one consisting of GR with DNMT3 for the methylation and the other the MeCP2, bound to methylated region, interact with HDAC1 to induce repression (68).

Clock Genes

The clock system and the HPA axis are interconnected regulatory systems. Cortisol circadian rhythm is modulated by the interaction between a central pacemaker, located in the hypothalamic suprachiasmatic nuclei, and the HPA axis (69). At the molecular level, mediators of the clock system and cortisol also communicate with each other, both acting as transcription factors of many genes to influence cellular functions.

In CS, the impact of chronic GC exposure on clock genes expression was recently evaluated using peripheral blood samples from patients with active disease compared with healthy subjects. The circadian rhythm of peripheral clock gene expression (CLOCK, BMAL, PER1-3, and CRY1) was abolished as a result of hypercortisolism, and that may contribute to metabolic disorders observed in Cushing patients (70). Another study, which investigated persistent changes induced by hypercortisolism in visceral adipose tissue, found that the expression of clock genes, such as PER1, remained altered in association with persistent epigenetic changes in both H3K4me3 and H3K27ac induced by hypercortisolism even after the resolution of hypercortisolism (12). This suggests that chronic exposure to GC may induce sustained epigenetic changes that can influence clock genes expression. Nevertheless, further studies are warranted to better elucidate how long-term exposure to GC impacts clock genes expression using the epigenetic machinery.

Glucocorticoid Effects on MicroRNAs

Along with histone modification and DNA methylation, microRNAs (miRNAs) have emerged as an epigenetic mechanism capable of impacting gene expression without changing DNA sequence (15). Interestingly, miRNA expression itself is also under the influence of epigenetic modifications through promoter methylation like any other protein-encoding genes (71).

MicroRNAs are small (about 20-25 nucleotides in length) non-coding RNAs that are important in transcriptional silencing of messenger RNA (mRNA). By partially pairing with mRNA, miRNAs can either induce mRNA degradation or inhibit mRNA translation to protein. MiRNAs regulate the translation of about 50% of the transcriptome, allowing them to play an important role in a wide range of biological functions, such as cell differentiation, proliferation, metabolism, and apoptosis under normal physiological and pathological situations. Some miRNAs can be classified as oncogenes or tumor suppressing genes, and aberrant expression of miRNAs may be implicated in tumor pathogenesis (71-73).

Insight into the regulation of miRNA expression is, therefore, crucial for a better understanding of tumor development and other human diseases, including cardiac, metabolic, and neurological disorders (73, 74). There are different regulatory mechanisms involved in miRNA expression, including transcriptional factors such as GR-GC. GC may modulate miRNA expression through direct binding to GRE in the promoter region of the host gene, as observed in hemopoietic tumor cells (75). In addition to transcriptional activation, in vascular smooth muscle cells, Dex treatment induces downregulation of DNMT1 and DNMT3a protein levels and reduces the methylation of miRNA-29c promoter, resulting in an increased expression of miRNA-29c (76). Interestingly, it was demonstrated that the increased expression of miRNA-29 family (miRNA-29a, -29b, and -29c) associates with metabolic dysfunction, such as obesity and insulin resistance, which pertains to CS (77, 78). With regard to metabolic dysfunction, miRNA-379 expression was shown to be upregulated by GC and its overexpression in the liver resulted in elevated levels of serum triglycerides associated with very low-density lipoprotein (VLDL) fraction in mice (79). In obese patients, the level of hepatic miRNA-379 expression was higher compared to nonobese patients and positively correlated with serum cortisol and triglycerides (79). Hence, GC-responsive miRNA may be, at least in part, a mediator to GC-driven metabolic conditions in CS.

In pathological conditions, such as seen in CS, prolonged exposure to an elevated cortisol level results in a wide range of comorbidities. It can be hypothesized that the chronic and excessive glucocorticoid levels may induce an aberrant miRNA expression that might impact several cellular processes related to bone and cardiometabolic disorders. A recent study addressed the impact of hypercortisolism on bone miRNA of patients with active CD compared to patients with nonfunctional pituitary adenomas. Significant changes in bone miRNA expression levels were observed, suggesting that the disruption of miRNA may be partially responsible for reduced bone formation and osteoblastogenesis (80). Similarly, altered expression levels of selected miRNAs related to endothelial biology in patients with CS may point to a contribution to a high incidence of cardiovascular disorders in Cushing patients (81). Therefore, dysregulated miRNAs as a consequence of high cortisol levels may underpin the development and progression of comorbidities related to CS. To the best of our knowledge, it is currently not clear whether miRNA dysregulation persists after resolution of hypercortisolism, thus contributing to the persistence of some comorbidities. This hypothesis needs to be further investigated.

MicroRNA can also be used as a diagnostic tool in CS. A study was performed to identify circulating miRNA as a biomarker to differentiate patients with CS from patients with suspected CS who had failed diagnostic tests (the control group) (82). It was observed that miRNA182-5p was differentially expressed in the CS cohort compared to the control group; therefore, it may be used as a biomarker (82). However, a large cohort is necessary to validate this finding (82). In corticotroph tumors, downregulation of miRNA 16-1 expression was observed relative to normal pituitary tissue (83). In contrast, the plasma level of miRNA16-5p was found to be significantly higher in CD compared to ectopic Cushing (EAS) and healthy controls (84). This finding suggests that miRNA16-5p may be a biomarker capable to differentiate the 2 forms of ACTH-dependent Cushing (84).

Epidrugs and Glucocorticoid Action in Cushing’s Syndrome

The interest in understanding the epigenetic mechanism of GC action in the context of CS is based on reversibility of epi-marks, such as DNA methylation and histone modifications, using epidrugs (85, 86). The biological characteristics of epigenetic drugs and their target have been extensively explored. Their effectiveness as antitumor drugs have been tested on corticotroph tumors using in vitro studies (87-89). However, a limited number of studies have explored the role of epidrugs as a therapeutic tool in reversing the genomic action of GC in CS, particularly in comorbidities induced by hypercortisolism (90, 91).

The use of histone deacetylase inhibitors (HDACi) may reduce the genomic action of GC (90-92). It has been demonstrated that the use of the HDAC inhibitor valproic acid increases the acetylation level of GR, consequently attenuating the genomic action of GC. In an experimental Cushing model in rats, the use of valproic acid decreased expression of genes related to lipogenesis, gluconeogenesis, and ion regulators in the kidney that ultimately reduces hepatic steatosis, hyperglycemia, and hypertension in ACTH-infused rats (90, 91).

More studies evaluating the effects of epidrugs influencing the GC actions are warranted to further elucidate the underlying mechanisms and to explore potential treatment modalities to reverse long-lasting consequences of chronic corticoid exposure.

Conclusions

In physiologic conditions, GC are secreted in pulses following a circadian rhythm pattern, as opposed to a constant, chronic, and high GC exposure in CS. This pathological pattern may account for numerous devastating effects observed in CS (7). Yet, the expressed genome in response to chronic GC exposure may potentially be abnormal, leading to dysregulation in clock genes, among other effects.

GC levels may return to a normal circadian pattern in response to a successful treatment, but with incomplete reversibility of some CS features, which may in part be explained by epigenetic changes. The epigenetic machinery is used by GC to induce dynamic changes in chromatin to modulate gene expression. (Fig. 2) It seems that most of chromatin modifications are reversible, but some may persist resulting in long-term epigenetic changes. (Table 1)

Table 1.

Evidence of interaction between glucocorticoid and epigenetic machinery

Epigenetic changes/epigenetic enzymes Action
Histone acetylation (HAT)

  • Glucocorticoid receptors (GR) recruit co-regulators, such as CREB binding protein (CBP), which has intrinsic histone acetyltransferase (HAT) activity that modifies the chromatin structure from an inactive to an active state (25, 33, 35).
Histone deacetylation (HDAC)

  • GR recruit histone deacetylases (HDACs) to turn chromatin less accessible and suppress gene transcription (25, 35).

  • The trans-repression process on POMC transcription achieved by glucocorticoids (GC) involves the histone deacetylation enzyme (HDAC2).

  • GC mediates the upregulation of HDAC2 in rats exposed to chronic stress (40).
Histone demethylase (JMJD3)

  • GC suppress transcription of JMJD3 in endothelial cells treated with TNFα (41).
Histone modifications

  • Using ChIP-seq, a study in mice treated for 5 weeks with corticosterone showed higher levels of histone modifications (H3K4me3, H3K27ac) compared to control mice. In mice after a 10-week washout period, persistence of this epigenetic fingerprint was observed, which was associated with long-lasting changes in gene expression (12).
DNA methylation (DNMT3B) and histone deacetylation (HDAC1)

  • GC mediates CRH downregulation through DNMT3B to the promoter in order to methylate a specific region and recruitment of chromatin modify corepressor HDAC (68).
DNA hypomethylation

  • GC induces downregulation of DNMT1 in AtT20 (mouse corticotroph adenoma cell line) (20).

  • GC induces upregulation of TET enzyme expression which was described in retinal and osteocyte cell line model (42, 43).

  • An experimental study in mice previously exposed to high levels of GC showed differentially methylated regions (DMR) induced by GC treatment, of which the majority was loss of the methylation (11).

  • Reduced DNA methylation in FKBP5 gene was found in patients in active disease and also in remission state of Cushing syndrome (CS) as compared to a healthy control group (10).

  • A genome-wide DNA methylation analysis showed a lower average of DNA methylation in patients in remission of CS compared to controls (9).

  • A study using whole blood methylation profile demonstrated an association between cortisol excess and DNA hypomethylation in patients with CS (46).
Open in new tab

Further studies are needed to elucidate how chronic exposure to GC leads to incomplete reversibility of CS morbidities via sustained modulation of the epigenetic machinery and possibly other mechanisms. Subsequent identification of therapeutic targets may offer new perspective for treatments, for example, with epidrugs, aiming to reverse hypercortisolism-related comorbidities.

Funding

The authors received no financial support for this manuscript.

Disclosures

T.P., R.A.F., and L.J.H. have nothing to declare.

Data Availability

Data sharing is not applicable to this article, as no datasets were generated or analyzed during the current study.

From https://academic.oup.com/jcem/advance-article/doi/10.1210/clinem/dgae151/7633538?searchresult=1

Filed under: Cushing's, symptoms | Tagged: , , , , , , | Leave a comment »

Most Subclinical Cushing’s Patients Don’t Require Glucocorticoids After Adrenalectomy

Posted on October 23, 2018 by MaryO

Patients with subclinical hypercortisolism, i.e., without symptoms of cortisol overproduction, and adrenal incidentalomas recover their hypothalamic-pituitary-adrenal (HPA) axis function after surgery faster than those with Cushing’s syndrome (CS), according to a study.

Moreover, the researchers found that an HPA function analysis conducted immediately after the surgical removal of adrenal incidentalomas — adrenal tumors discovered by chance in imaging tests — could identify patients in need of glucocorticoid replacement before discharge.

Using this approach, they found that most subclinical patients did not require treatment with hydrocortisone, a glucocorticoid taken to compensate for low levels of cortisol in the body, after surgery.

The study, “Alterations in hypothalamic-pituitary-adrenal function immediately after resection of adrenal adenomas in patients with Cushing’s syndrome and others with incidentalomas and subclinical hypercortisolism,” was published in Endocrine.

The HPA axis is the body’s central stress response system. The hypothalamus releases corticotropin-releasing hormone (CRH) that acts on the pituitary gland to release adrenocorticotropic hormone (ACTH), leading the adrenal gland to produce cortisol.

As the body’s defense mechanism to avoid excessive cortisol secretion, high cortisol levels alert the hypothalamus to stop producing CRH and the pituitary gland to stop making ACTH.

Therefore, in diseases associated with chronically elevated cortisol levels, such as Cushing’s syndrome and adrenal incidentalomas, there’s suppression of the HPA axis.

After an adrenalectomy, which is the surgical removal of one or both adrenal glands, patients often have low cortisol levels (hypocortisolism) and require glucocorticoid replacement therapy.

“Most studies addressing the peri-operative management of patients with adrenal hypercortisolism have reported that irrespective of how mild the hypercortisolism was, such patients were given glucocorticoids before, during and after adrenalectomy,” the researchers wrote.

Evidence also shows that, after surgery, glucocorticoid therapy is administered for months before attempting to test for recovery of HPA function.

For the past 30 years, researchers at the University Hospitals Cleveland Medical Center have withheld glucocorticoid therapy in the postoperative management of patients with ACTH-secreting pituitary adenomas until there’s proof of hypocortisolism.

“The approach offered us the opportunity to examine peri-operative hormonal alterations and demonstrate their importance in predicting need for replacement therapy, as well as future recurrences,” they said.

In this prospective observational study, the investigators extended their approach to patients with subclinical hypercortisolism.

“The primary goal of the study was to examine rapid alteration in HPA function in patients with presumably suppressed axis and appreciate the modulating impact of surgical stress in that setting,” they wrote. Collected data was used to decide whether to start glucocorticoid therapy.

The analysis included 14 patients with Cushing’s syndrome and 19 individuals with subclinical hypercortisolism and an adrenal incidentaloma. All participants had undergone surgical removal of a cortisol-secreting adrenal tumor.

“None of the patients received exogenous glucocorticoids during the year preceding their evaluation nor were they taking medications or had other illnesses that could influence HPA function or serum cortisol measurements,” the researchers noted.

Glucocorticoid therapy was not administered before or during surgery.

To evaluate HPA function, the clinical team took blood samples before and at one, two, four, six, and eight hours after the adrenalectomy to determine levels of plasma ACTH, serum cortisol, and dehydroepiandrosterone sulfate (DHEA-S) — a hormone produced by the adrenal glands.

Pre-surgery assessment of both groups showed that patients with an incidentaloma plus subclinical hypercortisolism had larger adrenal masses, higher ACTH, and DHEA-S levels, but less serum cortisol after adrenal function suppression testing with dexamethasone.

Dexamethasone is a man-made version of cortisol that, in a normal setting, makes the body produce less cortisol. But in patients with a suppressed HPA axis, cortisol levels remain high.

After the adrenalectomy, the ACTH concentrations in both groups of patients increased. This was found to be negatively correlated with pre-operative dexamethasone-suppressed cortisol levels.

Investigators reported that “serum DHEA-S levels in patients with Cushing’s syndrome declined further after adrenalectomy and were undetectable by the 8th postoperative hour,” while incidentaloma patients’ DHEA-S concentrations remained unchanged for the eight-hour postoperative period.

Eight hours after surgery, all Cushing’s syndrome patients had serum cortisol levels of less than 2 ug/dL, indicating suppressed HPA function. As a result, all of these patients required glucocorticoid therapy for several months to make up for HPA axis suppression.

“The decline in serum cortisol levels was slower and less steep [in the incidentaloma group] when compared to that observed in patients with Cushing’s syndrome. At the 6th–8th postoperative hours only 5/19 patients [26%] with subclinical hypercortisolism had serum cortisol levels at ≤3ug/dL and these 5 were started on hydrocortisone therapy,” the researchers wrote.

Replacement therapy in the subclinical hypercortisolism group was continued for up to four weeks.

Results suggest that patients with an incidentaloma plus subclinical hypercortisolism did not have an entirely suppressed HPA axis, as they were able to recover its function much faster than the CS group after surgical stress.

From https://cushingsdiseasenews.com/2018/10/11/most-subclinical-cushings-patients-dont-need-glucocorticoids-post-surgery-study/?utm_source=Cushing%27s+Disease+News&utm_campaign=a881a1593b-RSS_WEEKLY_EMAIL_CAMPAIGN&utm_medium=email&utm_term=0_ad0d802c5b-a881a1593b-72451321

Filed under: adrenal, Cushing's, pituitary | Tagged: , , , , , , | Leave a comment »

Blood Lipid Levels Linked to High Blood Pressure in Cushing’s Disease Patients

Posted on April 25, 2018 by MaryO

High lipid levels in the blood may lead to elevated blood pressure in patients with Cushing’s disease, a Chinese study shows.

The study, “Evaluation of Lipid Profile and Its Relationship with Blood Pressure in Patients with Cushing’s Disease,” appeared in the journal Endocrine Connections.

Patients with Cushing’s disease often have chronic hypertension, or high blood pressure, a condition that puts them at risk for cardiovascular disease. While the mechanisms of Cushing’s-related high blood pressure are not fully understood, researchers believe that high levels of cortisol lead to chronic hypertension through increased cardiac output, vascular resistance, and reactivity to blood vessel constrictors.

In children and adults with Cushing’s syndrome, the relationship between increased cortisol levels and higher blood pressure has also been reported. Patients with Cushing’s syndrome may remain hypertensive even after surgery to lower their cortisol levels, suggesting their hypertension is caused by changes in blood vessels.

Studies have shown that Cushing’s patients have certain changes, such as increased wall thickness, in small arteries. The renin-angiotensin system, which can be activated by glucocorticoids like cortisol, is a possible factor contributing to vascular changes by increasing the uptake of LDL-cholesterol (LDL-C) — the “bad” cholesterol — in vascular cells.

Prior research showed that lowering cholesterol levels could benefit patients with hypertension and normal lipid levels by decreasing the stiffness of large arteries. However, the link between blood lipids and hypertension in Cushing’s disease patients is largely unexplored.

The study included 84 patients (70 women) referred to a hospital in China for evaluation and diagnosis of Cushing’s disease. For each patient, researchers measured body mass index, blood pressure, lipid profile, and several other biomarkers of disease.

Patients with high LDL-cholesterol had higher body mass index, blood pressure, cholesterol, triglycerides, and apolipoproteinB (apoB), a potential indicator of atherosclerosis and cardiovascular disease.

Data further revealed an association between blood pressure and lipid profile, including cholesterol, triglycerides, apoB and LDL-c. “The results strongly suggested that CHO (cholesterol), LDL-c and apoB might predict hypertension more precisely in [Cushing’s disease],” the scientists wrote.

They further add that high cholesterol, LDL-cholesterol, and apoB might be contributing to high blood pressure by increasing vessel stiffness.

Additional analysis showed that patients with higher levels of “bad” cholesterol — 3.37 mmol/L or higher — had higher blood pressure. This finding remained true, even when patients were receiving statins to lower their cholesterol levels.

No association was found between blood pressure and plasma cortisol, UFC, adrenocorticotropic hormone, or glucose levels in Cushing’s disease patients.

These findings raise some questions on whether lipid-lowering treatment for high blood pressure and cardiovascular disease would be beneficial for Cushing’s disease patients. Further studies addressing this question are warranted.

Adapted from https://cushingsdiseasenews.com/2018/04/24/blood-pressure-linked-lipid-levels-cushings-disease-study/

Filed under: adrenal, Clinical trials, Cushing's, pituitary, symptoms | Tagged: , , , , , , , , | Leave a comment »

Cushing’s Patients at Risk for Autoimmune Diseases After Condition Is Resolved

Posted on March 8, 2018 by MaryO

Children with Cushing’s syndrome are at risk of developing new autoimmune and related disorders after being cured of the disease, a new study shows.

The study, “Incidence of Autoimmune and Related Disorders After Resolution of Endogenous Cushing Syndrome in Children,” was published in Hormone and Metabolic Research.

Patients with Cushing’s syndrome have excess levels of the hormone cortisol, a corticosteroid that inhibits the effects of the immune system. As a result, these patients are protected from autoimmune and related diseases. But it is not known if the risk rises after their disease is resolved.

To address this, researchers at the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) examined 127 children with Cushing’s syndrome at the National Institutes of Health from 1997 until 2017.

Among the participants, 77.5 percent had a pituitary tumor causing the disease, 21.7 percent had ACTH-independent disease, and one patient had ectopic Cushing’s syndrome. All patients underwent surgery to treat their symptoms.

After a mean follow-up of 31.2 months, 7.8 percent of patients developed a new autoimmune or related disorder.

Researchers found no significant differences in age at diagnosis, gender, cortisol levels, and urinary-free cortisol at diagnosis, when comparing those who developed autoimmune disorders with those who didn’t. However, those who developed an immune disorder had a significantly shorter symptom duration of Cushing’s syndrome.

This suggests that increased cortisol levels, even for a short period of time, may contribute to more reactivity of the immune system after treatment.

The new disorder was diagnosed, on average, 9.8 months after Cushing’s treatment. The disorders reported were celiac disease, psoriasis, Hashimoto thyroiditis, Graves disease, optic nerve inflammation, skin hypopigmentation/vitiligo, allergic rhinitis/asthma, and nerve cell damage of unknown origin responsive to glucocorticoids.

“Although the size of our cohort did not allow for comparison of the frequency with the general population, it seems that there was a higher frequency of optic neuritis than expected,” the researchers stated.

It is still unclear why autoimmune disorders tend to develop after Cushing’s resolution, but the researchers hypothesized it could be a consequence of the impact of glucocorticoids on the immune system.

Overall, the study shows that children with Cushing’s syndrome are at risk for autoimmune and related disorders after their condition is managed. “The presentation of new autoimmune diseases or recurrence of previously known autoimmune conditions should be considered when concerning symptoms arise,” the researchers stated.

Additional studies are warranted to further explore this link and improve care of this specific population.

From https://cushingsdiseasenews.com/2018/03/06/after-cushings-cured-autoimmune-disease-risk-looms-study/

Filed under: Cushing's, pituitary, Treatments | Tagged: , , , , , , , , , , , , , , , , , , | Leave a comment »

Next Page »