Elevated 4β‐hydroxycholesterol/cholesterol ratio in anorexia nervosa patients

Abstract Recent studies have shown that the cytochrome P450 (CYP) 3A phenotype marker 4β‐hydroxycholesterol/cholesterol (4βOHC/C) ratio is negatively correlated with body weight in healthy volunteers, and that obese patients have lower 4βOHC levels than healthy controls. However, 4βOHC/C ratio in underweight patients has yet to be reported. The aim of this study was to examine potential differences in CYP3A activity between underweight patients with anorexia nervosa and normal‐weight volunteers by measuring plasma 4βOHC/C ratio. Furthermore, we wished to describe any association between body mass index (BMI) and 4βOHC/C ratio in underweight patients. A total of 20 underweight patients and 16 normal‐weight volunteers were included in the study, all females. Underweight patients had a median 4βOHC/C ratio (molar ratio × 10−5) of 2.52 (range, 0.90–11.3) compared to 1.29 (0.56–2.09) in normal‐weight subjects (Mann‐Whitney P = 0.0005). 4βOHC/C ratio was negatively correlated with BMI in underweight patients (r = −0.56, P = 0.011), and in the whole study population (r = −0.67, P < 0.0001). This suggests that the negative correlation between 4βOHC/C and BMI, which has previously been reported between 4βOHC/C and body weight in healthy volunteers, extends to underweight patients. The findings indicate that CYP3A activity increases with decreasing BMI, resulting in higher CYP3A activity in underweight patients compared to normal‐weight subjects. The potential clinical relevance of this needs to be studied further by comparing pharmacokinetics of drugs subjected to CYP3A‐mediated metabolism in underweight vs. normal‐weight individuals.

inter-individual variability in CYP3A activity exists due to a combination of genetic, environmental, and endogenous factors. 2 CYP3A phenotype can be measured by the endogenous biomarker 4βhydroxycholesterol (4βOHC), which is metabolized from cholesterol by CYP3A4 and CYP3A5, the two most important CYP3A enzymes in humans. 3,4 Variations in cholesterol concentration have been found to explain about 10% of 4βOHC variation, 5 and 4βOHC/cholesterol (4βOHC/C) ratio is preferable to 4βOHC as CYP3A biomarker in patient groups where cholesterol levels are abnormal. 6 Recent studies have shown that body weight is negatively correlated with 4βOHC/C ratio in healthy volunteers, 7 and that obese patients have lower 4βOHC levels than healthy controls. 8 Studies on the clearance of a number of other CYP3A substrates suggest that CYP3A activity is reduced by 10-35% in obese patients. 9 Furthermore, Ulvestad et al. reported strong negative correlation between body mass index (BMI) and CYP3A protein expression in liver and intestines. 10 Altogether, this indicates that CYP3A activity decreases with increasing body weight. However, 4βOHC/C ratio in underweight patients is yet to be reported, and it is not known whether the correlation between BMI and CYP3A activity extends to underweight patients.
The aim of our study was to examine potential differences in CYP3A activity between underweight patients with anorexia nervosa and normal-weight volunteers by measuring plasma 4βOHC/C ratio. Furthermore, we wished to describe any association between BMI and 4βOHC/C ratio in underweight patients.

| Subjects
Patients with severe anorexia nervosa (n = 20) were included from an inpatient unit at the Regional Department for Eating Disorders, All subjects gave written, informed consent. The study was approved by the Regional Committee for Medical and Health Research Ethics South-East.

| 4βOHC analyses
Plasma concentration of 4βOHC was determined by a previously described UPLC-MS/MS assay. 11,12 The lower limit of quantification was 10 ng/mL. Intra-and interday imprecision and inaccuracy for the method were <15% at 10 ng/mL and <4% at 644 ng/mL (n = 6). 11 All samples were stored at −80°C between sampling and analysis. For the underweight patient samples, stability of 4βOHC during storage was ensured by controlling that 4αOHC concentration was lower than 4βOHC concentration. 13 This was not done for normal weight samples, since they were analyzed only 5 months after sampling; within the time frame that 4βOHC is known to maintain stability. 14

| Endpoints and statistical analyses
BMI was calculated as body weight divided by height squared (kg/ m 2 ). 4βOHC/C ratio was calculated as 4βOHC concentration (nmol/ L) divided by total cholesterol concentration (mmol × 10 6 /L).
Mann-Whitney U tests were used for comparisons of 4βOHC/C ratio and other characteristics between anorexia nervosa patients and normal-weight subjects. Spearman correlation was used to evaluate association between BMI and 4βOHC/C ratio and between 4βOHC and cholesterol. Statistical significance was considered as P < 0.05. GraphPad Prism for Windows, version 6.01 (GraphPad Software, La Jolla, CA), was used for statistical analyses and graphical presentations.

| RESULTS
Clinical and demographic characteristics of included subjects are listed in Table 1. All included subjects were female, and there were no differences in age or total cholesterol between the two groups (P > 0.1). None of the included subjects used CYP3A inducers according to the Flockhart CYP Drug Interaction Table. 15 One of the anorexia nervosa patients used fluoxetine, a CYP3A inhibitor. 15 The patient was not excluded since results showed higher 4βOHC concentration in underweight patients compared to normal-weight subjects. All drugs used by included subjects are listed in Table 2.
The association between 4βOHC and cholesterol concentration in the whole study population was significant (r = 0.41, P = 0.013).
Underweight patients had a median 4βOHC/C ratio (molar ratio ×  negatively correlated with BMI in underweight patients (r = −0.56, P = .011) (Figure 2A), and in the whole study population (r = −0.67, P < 0.0001) ( Figure 2B).  16 Cachectic patients have been reported to have both increased and decreased CYP3A metabolism, 17,18 but not to have altered liver content of CYP3A proteins. 19 Altogether, conflicting reports make it difficult to conclude whether CYP3A metabolism in underweight patients diverges from normal-weight subjects.

| DISCUSSION
In this study, we report that anorexia nervosa patients have higher 4βOHC concentration and 4βOHC/C ratio compared to normal-weight volunteers, and hypothesize that this reflects elevated CYP3A activity in underweight patients. Anorexia nervosa patients often have hypercholesterolemia. 20 Hyperadiponectinemia is known to occur in anorexic patients, 21 and may be related to increased cholesterol synthesis. 22 To account for impact of altered cholesterol levels on 4βOHC concentration, we consider 4βOHC/C ratio to be a more appropriate CYP3A biomarker in this study population. However, whether using 4βOHC concentration or 4βOHC/C ratio, we report approximately twice the biomarker level in underweight patients compared to normal-weight volunteers. The normal-weight volunteers had 4βOHC levels and 4βOHC/C ratio within the normal range. 23 Obesity has been associated with reduced plasma 4βOHC levels, 8 and hence emaciation could lead to increased 4βOHC levels.
However, distorted 4βOHC levels could be caused by changes both in production and elimination, and 4βOHC is further metabolized via CYP7A1 and CYP27A1. 3  In conclusion, we report that anorexia nervosa patients have twice the plasma 4βOHC/C ratio of normal-weight volunteers, and that 4βOHC/C ratio has a strong negative correlation with BMI. The findings indicate that CYP3A activity increases with decreasing BMI, resulting in higher CYP3A activity in underweight patients compared to normal-weight subjects. However, the potential clinical relevance needs to be studied further by comparing pharmacokinetics of drugs subjected to CYP3A-mediated metabolism in underweight vs.
normal-weight individuals.