Plasma oxysterol levels in luminal subtype breast cancer patients are associated with clinical data

Abstract

Oxygenated metabolites of cholesterol (oxysterols) have been previously demonstrated to contribute to progression of various cancers and to modulate resistance to breast cancer endocrine therapy in vitro. We measured prognostic roles of circulating levels of seven major oxysterols in the progression of luminal subtype breast carcinoma. Liquid chromatography coupled with tandem mass spectrometry was used for determination of levels of non-esterified 25-hydroxycholesterol, 27-hydroxycholesterol, 7α-hydroxycholesterol, 7-ketocholesterol, cholesterol-5α,6α-epoxide, cholesterol-5β,6β-exide, and cholestan-3β,5α,6β-triol in plasma samples collected from patients (n=58) before surgical removal of tumors. Oxysterol levels were then associated with clinical data of patients. All oxysterols except cholesterol-5α,6α-epoxide were detected in patient plasma samples. Circulating levels of 7α-hydroxycholesterol and 27-hydroxycholesterol were significantly lower in patients with small tumors (pT1) and cholesterol-5β,6β-epoxide and cholestan-3β,5α,6β-triol were lower in patients with stage IA disease compared to larger tumors or more advanced stages. Patients with higher than median cholestane-3β,5α,6β-triol levels had significantly worse disease-free survival than patients with lower levels (p=0.037 for all patients and p=0.015 for subgroup treated only with tamoxifen). In conclusion, this study shows, for the first time, that circulating levels of oxysterols, especially cholestane-3β,5α,6β-triol, may have prognostic roles in patients with luminal subtype breast cancer.

1. Introduction

Breast cancer (OMIM: 114480) is the most frequently diagnosed female cancer and the leading cause of cancer death among females worldwide (1). Despite significant progress in the early diagnosis and targeted therapy over the last several decades, a large number of patients experience treatment resistance and disease relapse. The discovery of biomarkers enabling better prognostication of these events and new therapy targets are needed for the treatment of these patients.

Among a plethora of newly reported circulating entities that could serve as prognostic and predictive biomarkers, oxygenated metabolites of cholesterol (termed oxysterols (2)) are of considerable interest. Oxysterols are formed in the human body or ingested in the diet and play important roles in the organism, e.g., interact with liver X receptors, retinoic acid receptor-related orphan receptors (RORs), estrogen receptors (ERs), glucocorticoid receptors (GRs) (3), oxysterol-binding proteins, drug transporters, the cholesterol signaling pathway (4), the hedgehog signaling pathway (5). Through modulation of these important cellular processes, oxysterols can contribute to several pathologies including carcinogenesis and cancer progression (6, 7).

Cholesterol has been demonstrated as a risk factor for the onset of breast cancer, and high levels have been associated with poor prognosis and recurrence, advocating for life-style and pharmacological interventions (8, 9). Since the finding that 27-hydroxycholesterol accelerates the growth and proliferation of mouse mammary gland tumors (10, 11), attention also turned to cholesterol metabolites. Several oxysterols are cytotoxic, promote cell death, block cell division, and interact with important signaling pathways (3, 12). Although two small-scale pilot studies reported changes in circulating blood levels of selected oxysterols in breast cancer patients before and during therapy (13, 14), suggesting their prognostic potential, a considerably larger study using homogeneous patient cohort is needed for a better understanding of this field.

In this study, levels of seven oxysterols (Fig. 1) were determined in plasma samples of patients with the luminal subtype (estrogen receptor-positive) breast cancer, collected before any therapy. Associations of these levels were made with clinical characteristics, including patient survival. Our observations provide new insight into the role of oxysterols in the biology of breast cancer.

Figure 1:

Steroidogenic pathway with oxysterols followed in this study.

Abbreviations: CHEH – cholesterol epoxide hydrolase, CH25H – cholesterol 25-hydroxylase, CYP – cytochrome P450, DHCR7 – dehydrocholesterol reductase, ROS – reactive oxygen species.

2. Materials and methods

2.1. Chemicals

25-Hydroxycholesterol (25HC, CAS No. 2140-46-7), 27-hydroxycholesterol (27HC, CAS No. 20380-11-4), 7-ketocholesterol (7KC, CAS No. 566-28-9), 7α-hydroxycholesterol (7αHC, CAS No. 566-26-7), cholesterol-5α,6α-epoxide (αEC, CAS No. 1250-95-9), cholesterol-5β,6β-epoxide (βEC, CAS No. 4025-59-6), and cholestane-3β,5α,6β-triol (CT, CAS No. 1253-84-5) and their deuterated standards (25-hydroxycholesterol-d₆, 27-hydroxycholesterol-d₆ 7-ketocholesterol-d₇, 7α-hydroxycholesterol-d₇, cholesterol-5α,6α-epoxide-d₇, cholesterol-5β,6β-epoxide-d₇, and cholestane-3ß,5α,6ß-triol-d₇) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL, USA). Methanol, isopropanol, and 2.6-di-tert-butyl-4-methylphenol (butylated hydroxytoluene) were purchased from Sigma-Aldrich (St. Louis, MO, USA), and HPLC-grade solvents were from Fisher Scientific (Pittsburgh, PA, USA).

2.2. Patients

A total of 58 patients with primary breast carcinoma was enrolled into the study. Patients were diagnosed in the period 2003 - 2011 at the Department of Oncology, Motol University Hospital in Prague. Blood samples were collected from each patient one day before surgical tumor removal. Patients were required to refrain from eating after 8 PM before the day of sampling, which was performed between 8 and 9 AM on the next day. Only incident patients without preoperative chemotherapy were included in the study. Breast carcinoma diagnosis, confirmed by histology examination of resection specimens, and signed informed consent by the patient were additional inclusion criteria.

Blood samples were collected in one vacutainer tube with anticoagulant potassium EDTA for plasma separation. The separation was performed immediately after blood withdrawal by centrifugation at 1,000 × g for 10 minutes at 4 °C. Then samples were divided into three separate aliquots, snap frozen, and stored at −80 °C until analysis. Diagnosis of all patients was confirmed histologically according to standard diagnostic procedures (15). Hormonal receptor expression was evaluated based on a 1% cut-off value. Immunohistochemistry was utilized for ERBB2 (erb-b2 receptor tyrosine kinase 2, OMIM: 164870) testing, with 3+ scores considered positive and 1+ negative. In the case of 2+ scores, fluorescent in situ hybridization (FISH) analysis was used for status confirmation. All patients had luminal subtype breast carcinoma (i.e., positive expression of estrogen receptor) and were treated with adjuvant endocrine therapy based on tamoxifen, aromatase inhibitors, or a combination of both. Personal and clinical characteristics of the patients, including adjuvant endocrine therapy, are summarized in Table 1. Disease-free survival (DFS) of patients was defined as the time elapsed between surgery and disease recurrence or death from any cause.

Table 1:

Clinical characteristics of patients

Characteristics	Patients, N (%)a
Age at diagnosis, mean ± SD (years)	62 ± 12
Menopausal status
Premenopausal	14 (24)
Postmenopausal	44 (76)
Tumor size (pT)
pT1	50 (86)
pT2	6 (10)
pT3	0 (0)
pT4	2 (4)
Lymph node metastasis (pN)
pN0	47 (87)
pN1	6 (11)
pN2	1 (2)
pNx	4–
Pathological stage
IA	43 (80)
IIA	9 (16)
IIB	1 (2)
IIIB	1 (2)
Unknown	4–
Histological type
Invasive ductal carcinoma	48 (83)
Other type (mostly lobular)	10 (17)
Pathological grade
G1	16 (28)
G2	39 (67)
G3	3 (5)
Estrogen receptor status
Positive	58 (100)
Negative	0 (0)
Progesterone receptor status
Positive	49 (84)
Negative	9 (16)
Expression of ERBB2
Positive	0 (0)
Negative	58 (100)
Adjuvant hormonal therapy
Tamoxifen alone	34 (59)
Tamoxifen with aromatase inhibitors	24 (41)

^aNumber of patients, with % in parentheses

All procedures performed in the present study were in accordance with the standards of the Ethics Committee for Multi-Centric Clinical Trials of the University Hospital Motol and 2^nd Faculty of Medicine, Charles University in Prague, Czech Republic (reference no. 17-28470A) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study.

2.3. Extraction of oxysterols from plasma samples

The method described by Helmschrodt et al. (16) was adapted for extraction of non-esterified oxysterols from plasma. Briefly, thawed plasma samples were centrifuged at 3,200 × g for 5 minutes and 80 μl of plasma was added to vials containing 80 μl of a methanolic solution of a mixture of internal standards (400 pg of the deuterated version of each oxysterol) or blank (methanol without oxysterols). Then 1.44 ml of a methanol/isopropanol (1:1, v/v) mixture was added, and the samples were thoroughly mixed using a vortex device for 1 min and centrifuged for 10 min at 12,500 × g. Supernatants were transferred into amber glass vials and evaporated to dryness at room temperature under a stream of nitrogen in the dark. Vials with dried extracts were filled with argon gas, sealed, and stored at −80 °C until analysis. On the day of analysis, extracts were reconstituted with 80 μl of a methanol/water mixture (3:1, v/v) and centrifuged prior to analysis for 5 min at 3,220 × g. All solutions contained 0.05% (v/v) butylated hydroxytoluene to prevent autoxidation (16).

2.4. Measurement of oxysterols

Dissolved extracts were subjected to the positive-ion atmospheric-pressure chemical ionization LC/MS/MS analysis as previously described (17). Briefly, oxysterols were resolved by UPLC (50 °C) on an Acquity BEH octadecylsilane (C18) column (1.7 μm: 1 mm × 100 mm) (Waters, Milford, MA, USA) using solvent mixtures of H₂O/CH₃OH/CH₃CN (30:3.5:66.5, v/v/v), holding for 1 min, then programming to CH₃OH/CH₃CN (5:95, v/v) over 4 min with a linear gradient at a flow rate of 0.16 ml/min and quantitated by positive-ion atmospheric-pressure chemical ionization MS/MS-MRM monitoring of product ions: m/z 367.2, 385.2 (25HC), m/z 367.2, 385.2 (27HC), m/z 367.2, 385.2 (7αHC), m/z 383.2, 401.2 (7KC), m/z 105.1, 367.5, 385.5 (αEC), m/z 105.1, 367.5, 385.5 (βEC), m/z 105.1, 367.5, 385.5 (CT), respectively, and m/z 373.2, (d₆-25HC), m/z 373.2, 391.2, (d₆-27HC), m/z 374.2, 392.2 (d₇-7αHC), m/z 390.2, 408.2 (d₇-7KC), m/z 105.2, 374.6, 392.6 (d₇-αEC), m/z 105.2, 374.6, 392.6 (d₇-βEC), m/z 105.2, 374.6, 392.6 (d₇-CT), respectively. The injection volume onto the column was 20 μl. Quantitation was done using an isotope ratio method with the internal standards. The limit of quantitation (LOQ) for oxysterols in human plasma was 0.05 ng per sample based on calibration curves for each oxysterol.

2.5. Data analysis

Raw data were first processed in order to identify outliers and deviations from normal distribution by the Shapiro-Wilk test. Differences between the compared groups of patients were evaluated by the independent Kruskal-Wallis or Spearman tests. For disease-free survival analysis, the Kaplan-Meier plot and the Breslow test were used to compare groups of patients divided by the median levels of oxysterols. A p-value of < 0.05 was considered statistically significant. Analyses were conducted by the statistical program SPSS v16.0 (SPSS, Chicago, IL, USA).

3. Results

3.1. Characteristics of patients

A total of 58 patients with luminal subtype of breast cancer and complete clinical follow up data was included in the study. The majority of patients was diagnosed with ductal adenocarcinoma histology at an early stage IA without axillary lymph nodes affected by metastasis (87% node negative). All patients had estrogen receptor positive and EBRB2 negative (luminal subtype) histology and received adjuvant therapy containing tamoxifen. About 41% of patients switched to aromatase inhibitors (Table 1). The median follow up time of patients was 106 months.

3.2. Levels of oxysterols in plasma from breast cancer patients

Oxysterols were analyzed in plasma samples of all patients. Plasma levels (means ± SD) of oxysterols in these patients are presented in Table 2. αEC was not detected in any of the analyzed samples due to the high background. Data were further processed by non-parametric tests due to the diversion of majority of data from the normal distribution (p < 0.05 by the Shapiro-Wilk test).

Table 2:

Circulating oxysterol levels in breast cancer patients

Oxysterol	Mean ± SD (ng/ml)	Range	Missing values (N)
25-Hydroxycholesterol	0.83 ± 0.62	0.07 – 3.16	17
27-Hydroxycholesterol	26.7 ± 43.1	3.48 – 231	0
7α-Hydroxycholesterol	7.62 ± 6.15	1.31 – 31.5	2
7-Ketocholesterol	6.68 ±4.74	0.27 – 16.2	1
Cholesterol-5α,6α-epoxide	not detected	not detected	58
Cholesterol-5β,6β-epoxide	15.8 ± 21.4	2.74 – 63.3	5
Cholestane-3β,5α,6β-triol	4.34 ± 2.38	0.98 – 12.5	6

Considerable inter-individual variability and several extreme values (outliers) were noted during detailed analysis of 27HC, 7αHC, and βEC (for value distribution see Fig. 2). Thus, separate statistical analyses with and without outliers were performed for these oxysterols.

Figure 2:

Distribution of levels of oxysterols in breast cancer patients.

The solid line represents the median and the dashed line the mean value. Outliers are highlighted with stars.

Correlation analysis indicated that 25HC was significantly correlated only with 7KC, while 27HC correlated with all oxysterols except 25HC. Additionally, 7KC levels were correlated with βEC and 7αHC levels were correlated with both βEC and CT (Table 3). In analysis without outliers all correlations were confirmed, except between 27HC and 7KC.

Table 3:

Correlations between levels of individual oxysterols in plasma of breast cancer patients

Oxysterola	25HC	27HC	7αHC	7KC	βEC	CT
25HC	-	- 0.073/0.649	0.192/0.241	0.512/0.001	−0.128/0.438	0.111/0.511
27HC		-	0.743/<0.001	- 0.409/0.002b	0.609/<0.001	0.505/<0.001
7αHC			-	- 0.053/0.702	0.573/<0.001	0.414/0.003
7KC				-	- 0.275/0.046	0.066/0.646
βEC					-	0.215/0.137
CT						-

^aSpearman’s correlation coefficient σ and p-value shown. Results confirmed by analysis without outliers are underlined

^bCorrelation between 27HC and 7KC was not significant in analysis without outliers (−0.171/0.268).

3.3. Associations of circulating oxysterol levels with clinical data of patients

The levels of 27HC and 7αHC were significantly lower in patients with small (pT1) tumors compared to patients with larger (pT2-4) tumors (Table 4). When outliers were removed, both associations still remained significant. Moreover, βEC and CT level was significantly lower in patients at an early stage IA compared to more advanced stages (confirmed for βEC without outliers). 7αHC and βEC were positively correlated with age in patients, both including and excluding outliers. CT levels were also positively correlated with age. Associations of 27HC levels with menopausal status, age, and disease stage and 7αHC levels with disease stage were not observed in analysis when outliers were excluded. In contrast, an association between tumor size and βEC was observed only after outliers were excluded (Table 4). Oxysterol levels were not associated with histological type and grade of tumor, presence of lymph node metastasis, or expression of progesterone receptor.

Table 4:

Differences in plasma levels of oxysterols between patients grouped by clinical data

Oxysterolsa
Clinical characteristics		25HC	27HC	7αHC	7KC	βEC	CT
Age at diagnosis		0.589	0.007	0.002	0.011	0.027	0.032
	Spearman’s correlation coefficient σ	- 0.087	0.348	0.405	0.938	0.301	0.298
	Without outliersb	-	0.112	0.006	-	0.026	-
	Spearman’s correlation coefficient σ	-	0.225	0.370	-	0.308	-
Menopausal status (pre vs. post)		0.866	0.044	0.068	0.778	0.053	0.143
	Without outliersb	-	0.181	0.102	-	0.079	-
Histology type (ductal vs. other)		0.822	0.450	0.181	0.578	0.733	0.722
Tumor grade (G1 vs. G2/G3)		0.933	0.321	0.219	0.873	0.199	0.479
Tumor size (pT1 vs. pT2-4)		0.376	0.011	0.007	0.077	0.072	0.089
	Without outliersb	-	0.017	0.003	-	0.038	-
Lymph node metastasis (pN01 vs. pN1)		0.790	0.202	0.528	0.581	0.108	0.152
Pathological stage (SIA vs.other)		0.580	0.007	0.035	0.392	0.015	0.018
	Without outliersb	-	0.143	0.182	-	0.022	-
Progesterone receptor (positive vs. negative)		0.531	0.957	0.907	0.526	0.917	0.214

^aA p-value by the Kruskal-Wallis or p-value and correlation coefficient for the Spearman test presented.

^bSignificant results were reanalyzed also without outliers. Confirmed significant results are underlined.

Associations between oxysterol levels and disease-free survival of patients treated with hormonal therapy were also evaluated. Patients with higher CT levels than median had significantly shorter disease-free survival compared with patients having lower levels than median (p = 0.037, Fig. 3A). The rest of the oxysterols did not associate with patient survival.

Figure 3:

Association between cholestane-3β,5α,6β-triol levels and the disease-free survival (DFS) of breast cancer patients.

Kaplan-Meier survival curves were plotted for the disease-free survival of patients divided into groups based upon median cholestane-3β,5α,6β-triol levels in their circulation. The solid line indicates the group with lower level and the dashed line the group with the higher level. A: all patients; B: patients treated with tamoxifen only.

At the end of the study, patients were stratified into subgroups treated with tamoxifen alone (n = 34) and those switched between tamoxifen and aromatase inhibitors in their therapy (n = 24). Both subgroups were then evaluated separately. The subgroups did not significantly differ in oxysterol levels. However, for the subgroup treated with tamoxifen only, CT levels were prognostic in the same manner as in unstratified patients (p = 0.015, Fig. 3B). The rest of the oxysterols were not significantly associated with disease-free survival in this subgroup. In the subgroup of patients treated with both tamoxifen and aromatase inhibitors, neither CT nor other oxysterols significantly predicted disease-free survival.

4. Discussion

Previous studies demonstrated that oxysterols can be detected in the circulation of breast cancer patients. One study reported significant differences in plasma profiles of several oxysterols in patients after the start of therapy with aromatase inhibitors (27HC and βEC) or tamoxifen (4β-hydroxycholesterol, 24-hydroxycholesterol, 25HC, and 7αHC), suggesting treatment-specific predictive role of these oxysterols (13). Another published article reported a significant increase of 7KC level after tumor removal and the beginning of adjuvant therapy (14). A very recent publication reports the presence of side chain oxysterols (24S-hydroxycholesterol, 25HC, and 27HC) in breast tumor tissues (18). The present follow-up study raised the question of whether selected oxysterols may have a prognostic role in a considerably larger and highly homogeneous group of breast cancer patients with luminal subtype of disease treated with hormonal therapy.

We provide the first evidence of association of CT levels in the plasma with prognosis of breast cancer patients. This association was observed both in all patients and in a subgroup treated solely with selective antiestrogen modulator tamoxifen, but not in the subgroup of patients treated with both tamoxifen and aromatase inhibitors.

CT is formed enzymatically by cholesterol epoxide hydrolase (ChEH, EC 3.3.2.11) from αEC and βEC (19, Fig. 1). CT is cytotoxic and promotes apoptosis in a human mammary carcinoma model MDA-MB-231 cell line in vitro, suggesting its therapeutic potential (12). Tumor suppressing effects of CT were also reported using both in vitro and in vivo models of prostate cancer, where treatment downregulated expression of proteins associated with epithelial-mesenchymal transition and focal adhesion (20). Moreover, CT is extensively metabolized by 11β-hydroxysteroid dehydrogenase type 2 (HSD11B2, EC 1.1.1.146) into 6-oxocholestan-3β,5α-diol (OCDO), a strong oncometabolite in breast cancer models both in vitro and in vivo (21). Thus, the published data in different models and experimental settings suggest that CT is a promising lead molecule for design of anticancer therapeutics. Our data show the opposite trend, i.e. high CT levels in plasma of breast cancer patients associated with shorter disease-free survival during or after hormonal therapy (Fig. 3B), i.e. indicating a poor prognosis of patients with estrogen receptorexpressing tumors. Whether CT has a similar effect in other histological subtypes and patients at more advanced stages of disease should be examined. Functional studies aimed at investigation of mechanism of action of candidate oxysterols in various in vitro and in vivo breast cancer models should also follow to confirm causality before potential translation of results into clinical practice.

An association of CT and βEC with disease stage and patients age was also noted in the present study, apparently for the first time as well as associations of 27HC and 7αHC with tumor size. Moreover, these oxysterols are highly correlated with each other and thus observed associations may be result of more general, so far unknown, phenomena that should be studied in more detail. Correlative data in homogeneous groups of cancer patients are virtually missing due to size constraints. Regards to the well-known prognostic importance of age and disease stage in breast cancer the role of association of CT with the axis age – stage – survival observed here needs further attention.

In contrary to a previously reported endogenous antiestrogenic role of 27HC (22) and higher levels of 25HC in metastatic compared to locally advanced breast cancer patients (13), side chain oxysterols were not associated with prognosis in the patients here. In agreement with our results, a recent study failed to validate a presumed protective effect (regarding risk of recurrence) of serum 27HC level manipulation in breast cancer patients treated with atorvastatin (23). Thus, our present data cannot add more rationale to the alternative approach of 27HC reduction in blood of breast cancer patients using vitamin D supplementation (24).

Compared to the published data on females (13, 25), the concentrations of side chain oxysterols measured in the present study were much lower. However, previous studies used alkaline hydrolysis and GC-MS analysis and thus determined both esterified and non-esterified oxysterols, whereas we used direct extraction with LC/MS/MS analysis and report levels of non-esterified or ‘free’ molecules. On the other hand, the 7KC, βEC, and CT levels we found are comparable to results published elsewhere (13, 25). We are currently modifying methods for αEC detection in these samples. Taken together, due to the technical difficulties associated with analysis of oxysterol mixtures from biological samples, e.g., considering their instability, sensitivity to material handling and storage, technical complexity (26, 27) and high variability among laboratories (28), results obtained with the patients we studied should be interpreted cautiously.

Certain limitations are inherent in our study. First, the modest sample size may be considered low, although this caveat applies more to subgroup analyses than to the whole set of results, in that our study analyzed the largest series of cancer patients by far. Second, oxysterols are highly unstable and prone to auto-oxidation. Long-term storage, even at ultra-low temperatures (−80 °C), has to be considered as a confounder that potentially could influence some results. It is therefore necessary to conduct further studies to validate the present results.

On the other hand, this study has a number of strengths. All patients were recruited and followed in one regional center and represent the most homogeneous group of breast cancer patients (early stage luminal subtype treated with hormonal therapy only) studied in this area to date. Moreover, the present study included patients followed for a very long period, a condition that is difficult to achieve.

In conclusion, the present study showed that plasma CT levels are associated with disease-free survival of breast cancer patients with luminal subtype treated with hormonal therapy. This observation is in line with the recently proposed central role of CT in a new cholesterol-metabolism related oncopathway (27). In future, prospective studies are necessary to further assess the robustness and clinical utility of this observation.

Highlights.

• oxygenated cholesterol metabolites (oxysterols) can contribute to cancer progression

• circulating levels of 7 oxysterols were assessed in luminal breast carcinoma patients

• high cholestane-3β,5α,6β-triol levels were associated with poor disease-free survival

• these observations provide new insight into the role of oxysterols in breast cancer

Abbreviations

APCI

atmospheric-pressure chemical ionization

αEC

cholesterol-5α,6α-epoxide

βEC

cholesterol-5β,6β-epoxide

ChEH

cholesterol epoxide hydrolase

CT

cholestane-3β,5α,6β-triol

DFS

disease-free survival

EC

Enzyme Commission

ER

estrogen receptor

ERBB2

erb-b2 receptor tyrosine kinase 2

FISH

fluorescent in situ hybridization

GR

glucocorticoid receptor

7αHC

7α-hydroxycholesterol

25HC

25-hydroxycholesterol

27HC

27-hydroxycholesterol

HSD11B2

11β-hydroxysteroid dehydrogenase type 2

7KC

7-ketocholesterol

LOQ

limit of quantitation

OCDO

6-oxocholestan-3β,5α-diol

OMIM

Online Mendelian Inheritance in Man

pT

pathological tumor size

pN

pathological lymph node metastasis

ROR

retinoic acid receptor-related orphan receptor