HCV Infection and Liver Cirrhosis Are Associated with a Less-Favorable Serum Cholesteryl Ester Profile Which Improves through the Successful Treatment of HCV

Background: Infection with hepatitis C virus (HCV) lowers serum cholesterol levels, which rapidly recover during therapy with direct-acting antivirals (DAAs). Serum cholesterol is also reduced in patients with liver cirrhosis. Studies investigating serum cholesterol in patients with chronic liver diseases are generally based on enzymatic assays providing total cholesterol levels. Hence, these studies do not account for the individual cholesteryl ester (CE) species, which have different properties according to acyl chain length and desaturation. Methods: Free cholesterol (FC) and 15 CE species were quantified by flow injection analysis high-resolution Fourier Transform mass spectrometry (FIA-FTMS) in the serum of 178 patients with chronic HCV before therapy and during treatment with DAAs. Results: Serum CEs were low in HCV patients with liver cirrhosis and, compared to patients without cirrhosis, proportions of CE 16:0 and 16:1 were higher whereas % CE 20:4 and 20:5 were reduced. FC levels were unchanged, and the CE/FC ratio was consequently low in cirrhosis. FC and CEs did not correlate with viral load. Four CE species were reduced in genotype 3 compared to genotype 1-infected patients. During DAA therapy, 9 of the 15 measured CE species, and the CE/FC ratio, increased. Relative to total CE levels, % CE 16:0 declined and % CE 18:3 was higher at therapy end. At this time, % CE 14:0, 16:0 and 16:1 were higher and % CE 20:4 and 22:6 were lower in the cirrhosis than the non-cirrhosis patients. Viral genotype associated changes of CEs disappeared at therapy end. Conclusions: The serum CE composition differs between patients with and without liver cirrhosis, and changes through the efficient elimination of HCV. Overall, HCV infection and cirrhosis are associated with a higher proportion of CE species with a lower number of carbon atoms and double bonds, reflecting a less-favorable CE profile.


Introduction
Chronic infection with hepatitis C virus (HCV) is a common cause of liver fibrosis, which may progress to cirrhosis [1]. Metabolic diseases such as liver steatosis or diabetes mellitus occur more often in association with chronic HCV infection. Fatty liver and diabetes can be direct effects of viral infection or a secondary event of insulin resistance due to viral infection. HCV patients may also suffer from the metabolic syndrome, which affects about 30% of the normal population. Thus, it is difficult to distinguish between these different causes of liver steatosis and insulin resistance. Hypertension and obesity are indicators of metabolic diseases, and HCV genotype 3 infection in patients with normal body score of 12, a low cholesterol esterification ratio was a good predictor of mortality [35]. A separate study calculated the CE/FC ratio but could not identify a difference between patients with predominantly alcoholic liver cirrhosis and liver-healthy controls [36].
Liver cirrhosis as well as HCV infection are characterized by low serum cholesterol [16,[30][31][32]. Whether FC and the different CE species are all similarly reduced has not been analyzed in great detail so far. FC has unique properties and high levels are cytotoxic [37]. An association of serum FC with bilirubin has been described in patients with alcoholic liver cirrhosis [38]. Notably, serum CE levels were not associated with markers of liver cirrhosis in this cohort [38].
The effective therapy of HCV elevates serum LDL and cholesterol levels, which are risk factors for cardiovascular diseases [39,40]. The elimination of HCV is, however, associated with a lower risk for cardiovascular diseases [3]. Not all of the CE species seem to contribute to an increased risk, and a higher polyunsaturated-to-saturated CE ratio protected from atherosclerosis [39]. Among the lipids with a strong predictive value for cardiovascular diseases were CE species with a low carbon number and a low double-bond content [41].
The main aim of the present analysis was to study the effect of DAA therapy on serum CE composition. It was also analyzed whether the CE profile differs between HCV patients with and without liver cirrhosis. Therefore, FC and 15 CE species were measured in the serum of HCV patients before therapy, at 4 weeks after treatment start-the time point where LDL levels have been recovered [8,18]-and at therapy end.

Study Cohort
Patients' sera were collected at the Department of Internal Medicine I (University Hospital of Regensburg) from October 2014 to September 2019 [42]. All of the 178 patients asked to participate in the study were suited for therapy with DAAs in agreement with HCV treatment guidelines [4] and all of them finished the study. None of the patients had been treated for HCV before. The patients were older than 18 years and were not coinfected with HBV or HIV.
Very few patients were using statins, and due to the potential interaction of DAA therapy with statins [43], the cholesterol-lowering drugs were paused during therapy.
Cirrhosis diagnosis by ultrasound was established based on nodular liver surface, small liver size and heterogeneous liver parenchyma [44]. Cut-off values used for the FIB-4 scoring were: >3.25: advanced fibrosis, <1.3: no fibrosis for patients younger than 65 years and <2: no fibrosis for patients older than 65 years [45].
Laboratory values were obtained from the Institute of Clinical Chemistry and Laboratory Medicine (University Hospital Regensburg). Individual laboratory values of ≥95% of the patients were available. Laboratory parameters of this study group have been published in an open access journal [46].

Measurement of FC and CEs
FC and CE were analyzed by flow injection analysis high-resolution Fourier Transform mass spectrometry (FIA-FTMS) on a QExactive Orbitrap (Thermo Fisher Scientific, Bremen, Germany), as described previously [19]. Multiplexed acquisition (MSX) was applied for the [M+NH4]+ of FC and the corresponding internal standard (FC[D7]). CE was recorded as [M+NH4]+ in positive ion mode in range m/z 500-1000 at a target resolution of 140,000 (at m/z 200). The accurate quantification of CE species by FIA-FTMS requires correction by individual response factors. The differences in response between CE species are because of structural features such as the length and double-bond number of acyl chains. The calculation of these CE species-specific response factors has been described in detail [19]. CE 17:0 and CE 22:0 (Sigma-Aldrich (Taufkirchen, Germany) were added as internal standards.

Statistical Analysis
Data are shown as boxes, and the mean value ± standard deviation is given. Boxplots (minimum, maximum, median, first and third quartiles, small circles or asterisks above or below the boxes mark outliers) were used in the case that only one lipid/lipid ratio was shown.
The Mann-Whitney U-test, one-way ANOVA, Kruskal-Wallis-test or t-test were used (SPSS Statistics 25.0 program, IBM, Armonk, New York, NY, USA; and Microsoft Excel 2016, Redmond, Washington, DC, USA). A value of p < 0.05 after adjusting for multiple comparisons was regarded as significant.

Association of Age, Gender, Body Mass Index, Liver Steatosis and Diabetes with Serum FC and CE Species
One hundred seventy-eight patients with chronic HCV were included in the study, and fifteen different cholesteryl ester (CE) species as well as free cholesterol (FC) were measured in serum. Levels of these lipids did not differ between the 74 females and 104 males and did not correlate with age or body mass index (BMI) (see Supplementary Figure S1A and Tables S1 and S2, Additional File 1). Accordingly, CE species and FC did not change with increasing body mass index (see Supplementary Figure S1B and Table S2, Additional File 1). The 74 patients with liver steatosis and the 104 patients without liver steatosis had similar concentrations of these CE species and FC (see Supplementary Figure S1C and Table S2, Additional File 1). Notably, levels of CE 20:5, 22:5 and 22:6 were reduced in the serum of the 20 diabetic patients ( Figure 1A), and the relative concentrations (% of total CEs) of these CE species declined (p < 0.05). FC levels were not changed in HCV patients with diabetes (Supplementary Table S2, Additional File 1), and thus, the CE/FC ratio was reduced ( Figure 1B).

CE Species Levels and Distribution in Patients with Advanced Liver Disease
There were 8 patients with diabetes in the group of HCV patients without cirrhosis, and 12 in the group of patients with cirrhosis. CE species did not differ between the nondiabetic and diabetic patients in these subgroups (data not shown). However, the number of patients was quite small and an effect of diabetes on serum CE levels cannot be ruled out. Thus, for the calculations of associations of CE species with measures of liver disease severity, HCV patients with diabetes were excluded.
FC was similar between patients with and without liver fibrosis/cirrhosis (p > 0.05). The CE/FC ratio was reduced in patients with a high FIB-4 score in comparison to the two groups with lower scores, and in patients with cirrhosis in comparison to non-cirrhosis patients ( Figure 3A,B).

Correlation of CE Species with the MELD Score and Laboratory Measures
In the patients without diabetes and without liver cirrhosis, there was a modest negative correlation of total CE levels with the MELD score ( Figure 4A; r = −0.267, p = 0.032). However, none of the single CE species correlated with the MELD score (Table 1). CE 18:2 was negatively linked with international normalized ratio (INR). Albumin, ALT, AST, creatinine, bilirubin, CRP and leukocytes did not correlate with any of the CE species (Table 1).    (Table 2). Table 2. Spearman correlation coefficients of CE species with the MELD score as well as laboratory measures of patients without diabetes and with liver cirrhosis before therapy. * p < 0.05, ** p < 0.01, *** p < 0.001.

Effect of DAA Therapy for HCV on Serum CE Species
The rise in total CE levels at 4 and 12 weeks after the start of DAA therapy was noticed in patients without and with liver cirrhosis ( Figure 5A,B). FC did not increase during therapy in the two subgroups ( Figure 5A,B). The CE/FC ratio was higher at 4 and 12 weeks after therapy start in both groups ( Figure 5C,D). Notably, the CE/FC ratio at 4 and 12 weeks after therapy start was still higher in patients without cirrhosis than those with liver cirrhosis (Figure 5C,D).  Figure 6A). The ratio of CE 18:0/16:0 is a marker of elongation of long-chain fatty acids family member 6 (ELOVL6) activity. The ratio CE 18 to CE 16 was modestly higher at therapy end (p = 0.02). % CE 16:0 was lower at 4 and 12 weeks after therapy start and % CE 18:3 was higher at these time points when compared to pre-treatment levels. The rise in % CE20:3 occurred at week 4 after therapy start and returned to pre-treatment levels at therapy end ( Figure 6B). None of the CE species were significantly increased at 4 or 12 weeks after therapy start in the cirrhosis group ( Figure 6C).   (Table 3). Table 3. CE levels in serum of patients with a low and a high FIB-4 score at therapy end. Only species which significantly differed between the two groups are listed. ** p < 0.01, *** p < 0.001.  (Table 4). Table 4. CE levels in serum of patients without and with liver cirrhosis diagnosed by ultrasound at therapy end. Only species which significantly differed between the two groups are listed. * p < 0.05, ** p < 0.01, *** p < 0.001. Notably, the proportion of CE 14:0, 16:0 and 16:1 was higher and that of CE 20:3, 20:4 and 22:6 was lower in patients with a high FIB4 score compared to patients with a low score ( Figure 7A). This suggests that fatty acid elongation may be also impaired. The ratio of CE 18:0/16:0 was lower in patients with a high FIB-4 score compared to those with a low score ( Figure 7B

Association of CE Species with Viral Titer and Genotype
CE species and FC did not correlate with viral load in the whole cohort, and in the HCV patients without and with ultrasound-diagnosed cirrhosis (Supplementary Table S3, Additional File 1 and data not shown).
Associations with genotype were calculated after the exclusion of diabetic patients and patients with liver cirrhosis. Forty-one patients had genotype 1a, forty-eight 1b, twentyeight 3a and thirteen had a different genotype and were grouped together (rare). Here, genotype 3a infection led to lower CE 15:0 and 16:0 compared to 1a, and lower CE 16:0, 18:2 and 20:4 compared to 1b (Figure 8). The relative content of CEs did not differ between the genotypes with the exception of CE15:0, which was lower in genotype 3a compared to 1a (p = 0.02). These genotype-associated variations disappeared at therapy end (data not shown).

Correlation of CE Species with HDL and LDL
In the group of HCV patients without liver cirrhosis, all but CE 22:4 positively correlated with LDL before therapy.  (Table 5). Table 5. Spearman correlation coefficients for the association of CE species with LDL and HDL before therapy and at therapy end in patients with and without ultrasound-diagnosed liver cirrhosis. * p < 0.05, ** p < 0.01, *** p < 0.001.

Discussion
This study shows that DAA therapy increases serum CE levels and leads to a more favorable CE profile. Patients with liver cirrhosis display an adverse CE profile, the relative content of saturated and monounsaturated CEs with short acyl chains is high and that of polyunsaturated CE species with 20 or 22 carbon atoms is low, and this does not change by DAA therapy.
LCAT forms CE 20:4, 22:5 and 22:6, whereas CE 16:0, 18:1 and 18:3 are derived from liver ACAT [23][24][25] (Figure 9). This might indicate that ACAT as well as LCAT activity are higher at therapy end-a hypothesis that has to be yet confirmed. However, various other enzymes and pathways have a role in CE composition. Cholesterol-esterifying activity in serum was found to be decreased in patients with acute hepatitis and chronic alcoholic liver disease, and besides LCAT, CE hydrolase seems to have a function herein [51]. Thus, it is currently unclear which pathways are involved. Notably, % CE 16:0 was reduced and % CE 18:3 was induced at therapy end, suggesting that a more favorable CE profile exists when HCV is efficiently cleared. Figure 9. Pathways which may contribute to the altered CE profile in HCV and liver cirrhosis. Free cholesterol in the liver is converted to CEs by ACAT, which produces atherogenic CE species. These are released from the liver by VLDL, which is converted to LDL in the circulation. Elongation of very-long chain fatty acids (ELOVL) 6 was found to belower expressed in the fibrotic liver, and activity of further elongases and desaturases may be impaired. Free fatty acids (FFA) are used for the synthesis of various lipid classes such as triglycerides, diacylglycerols, cholesteryl ester and phosphatidylcholine. LCAT in serum uses phosphatidylcholine to esterify cholesterol derived from peripheral cells and tissues. Here, highly unsaturated CE species are being formed. The CE species (found in the current work to be low when cirrhosis was defined by ultrasound), lipoproteins and enzymes which are reduced in cirrhosis are marked with a black arrow. A red arrow marks CE species and lipoproteins, which are low in HCV infection. Current findings are in agreement with improved ELOVL6, ACAT and LCAT activity at the end of DAA therapy. This is in line with a higher CE/FC ratio at therapy end. Whether the levels of these enzymes are reduced or FFA availability is limited must be experimentally clarified. ↓ Lower in cirrhosis, ↓ Lower in HCV. HCV patients have a higher incidence of metabolic diseases, including atherosclerosis [2]. There is evidence that cardiovascular diseases and insulin resistance improve after SVR by DAA therapy [52]. Future studies have to evaluate whether the decline of % CE 16:0 and the higher abundance of % CE 18:3 contribute to these beneficial effects.
The recovery of the CE species in the cirrhosis group was not significant. Total CE levels similarly increased during therapy in cirrhosis and non-cirrhosis patients, and were 119% and 116% higher at therapy end, respectively. This suggests that the elimination of HCV changes the CE levels of both groups. Because of the lower number of patients with cirrhosis, this was not significant in this subgroup.
Viral load did not correlate with FC or any of the CE species. Interestingly, genotype 3 infection reduced CE 15:0 and 16:0 in comparison to 1a infection, and CE 16:0, 18:2 and 20:4 in comparison to 1b infection. These differences did not persist at therapy end, showing that genotype 3 differently affects CE composition in comparison to genotype 1. Genotype 3-infected patients were described to have lower serum LDL cholesterol than genotype 1-infected patients [53]. Yet, CEs such as cholesteryl linoleate were higher in genotype 3 than genotype 1 [53]. The present analysis revealed lower levels of CE 18:2 in genotype 3-than 1b-infected patients, and currently, there is no explanation for these divergent results. The two studies agree that the genotype-related changes of the CE lipidome are not apparent in patients who achieved SVR [53].
Besides having low levels of serum cholesterol, patients with liver cirrhosis have a reduced CE/FC ratio. This may be due to a lower activity of enzymes involved in FC esterification. Cholesterol esterification in plasma has been described as a marker for liver function in patients with advanced stages of liver disease [35]. Interestingly, the CE/FC ratio increased during DAA treatment in the cirrhosis and non-cirrhosis group, showing that the cholesterol esterification rate is higher in both cohorts. Liver function does not greatly improve during DAA therapy [7,8] and the MELD score of our patient group did not change [46], indicating that the cholesterol esterification rate is not solely affected by liver disease severity.
Notably, serum FC levels did not differ between cirrhosis and non-cirrhosis patients before and after therapy. Pathways contributing to serum FC levels such as reverse cholesterol export [54] are, thus, not grossly impaired in cirrhosis.
Shorter CEs are the product of ACAT and the longer CEs from LCAT. Because CE 18:1 as well as CE 22:5 were low in cirrhosis, both cholesterol esterification pathways may be impaired ( Figure 9). Decreased LCAT activity has been described in patients with liver cirrhosis [29], but whether ACAT2 activity is also low has not been clarified yet. In addition, ELOVL6, which catalyzes the elongation of saturated and monounsaturated fatty acids with 12 to 16 carbon atoms, was found to be lower expressed in patients with advanced fibrosis [55] (Figure 9). Impaired activity of this enzyme may partly explain the depletion of longer chain fatty acids in CEs, and the CE 18-/C 16 ratio is low in our HCV patients with liver cirrhosis. Notably, the CE 18/CE 16 ratio was modestly higher at the end of DAA therapy, suggesting that HCV infection may also lower the activity of ELOVL6 (Figure 9). The different CE species are, however, not reduced to a similar extent in the serum of patients with liver cirrhosis. The proportion of CEs with short acyl chains and no or one double bond increased, whereas CEs with longer acyl chains and at least two double bonds declined. This did not change at therapy end, showing that liver cirrhosis is associated with a unfavorable CE profile. In patients with decompensated liver cirrhosis mostly because of alcohol abuse, the relative content of total plasma CE 14:0, 16:0 and 18:1 was higher and the relative content of CE 18:2 and 20:4 was lower compared to controls [56], and thus, were similarly altered, as observed in the HCV cohort studied herein. The activity of delta-6-desaturase, an enzyme involved in the synthesis of polyunsaturated fatty acids, was found to be reduced in liver cirrhosis [57], and this may also partly explain this observation.
Liver cirrhosis is not associated with a higher risk for atherosclerosis [58] and the pathophysiological role of the altered CE profile has still to be evaluated. A frequent comorbidity of liver cirrhosis is type 2 diabetes [59], but the associations of serum CE species with glucose homeostasis have not been finally resolved.
HCV-infected patients with diabetes studied herein had lower serum levels of CE 20:5, 22:5 and 22:6, and % CE 20:5 and 22:6 were reduced. In patients with type 2 diabetes, the proportions of CE 18:0 and CE 20:3n−6 were higher, and those of CE 18:1n−7 and C20:4n−6 were reduced compared to patients with normal glucose metabolism [60]. A further study identified protective associations of CE 18:1n−7 and 18:1n−9 and harmful associations of CE 18:3n−6 and 18:0 with insulin sensitivity and beta-cell function [61]. Current findings on the association of CE species with diabetes are inconsistent and further research is needed. It has to be noted that-in the current analysis-gender, BMI and age were not associated with FC levels or the change in any of the CE species analyzed.
In serum, CE species are carried by LDL, HDL and VLDL. LDL cholesterol content is about 2-fold higher than that of VLDL and HDL [20]. CE levels in serum mainly correlated with LDL levels, and thus, serum CE content was mostly related to LDL rather than HDL levels.
Interestingly, CE composition did not greatly differ between HDL, LDL and VLDL of healthy volunteers [20] and patients with liver cirrhosis [56]. This suggests that the CE profile of LDL, HDL and VLDL is changed in cirrhosis. How and whether an altered CE c influences the function of lipoproteins needs future investigations.
This study has limitations. Total serum CE and FC levels were measured but the composition of individual lipoproteins and LCAT activity were not analyzed. Serum was not collected in the fasted state, and dietary habits of the patients were not documented. However, serum cholesterol levels do not greatly vary during the day [62]. It is, moreover, very unlikely that the identified changes in CE composition achieved by HCV therapy and in patients with liver cirrhosis are explained by different diets. A further limitation is that healthy controls and patients with liver diseases of distinct etiologies were not included.
In summary, the present analysis showed that viral genotype, DAA therapy and cirrhosis differentially affect serum CE species levels. Prospective studies have to evaluate the prognostic value of CE species for cardiovascular diseases, insulin resistance, morbidity and mortality in chronic HCV infection and liver cirrhosis.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biomedicines10123152/s1, Figure S1: Serum cholesteryl ester (CE) species in relation to gender, body mass index (BMI) and steatosis in patients with chronic HCV. Table S1 Spearman correlation coefficients of the association of CE species and FC with age and BMI in the whole cohort before therapy. Table S2: Concentration of FC (nmol/ml) in relation to gender, BMI, liver steatosis and diabetes of patients with HCV before therapy start. Table S3: Spearman correlation coefficients of the association of FC and CE species with viral load before therapy.

Institutional Review Board Statement:
The study protocol was approved by the ethical committee of the University Hospital of Regensburg (14-101-0049) and the study was performed according to the updated guidelines of good clinical practice and updated Declaration of Helsinki.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The datasets generated and/or analyzed during the current study are available from the corresponding author on request.