Plasma Biomarker of Dietary Phytosterol Intake
Plasma Biomarker of Dietary Phytosterol Intake
Xiaobo Lin 0 1
Susan B. Racette 0 1
Lina Ma 0 1
Michael Wallendorf 0 1
Catherine Anderson Spearie 0 1
Richard E. Ostlund 0 1
Jr 0 1
0 1 Division of Endocrinology, Metabolism & Lipid Research, Department of Medicine, Institute for Clinical and Translational Sciences, Washington University School of Medicine , 660 South Euclid Ave., St. Louis, MO, 63110 , United States of America, 2 Program in Physical Therapy, Institute for Clinical and Translational Sciences, Washington University School of Medicine , 660 South Euclid Ave., St. Louis, MO, 63110 , United States of America, 3 Division of Biostatistics, Institute for Clinical and Translational Sciences, Washington University School of Medicine , 660 South Euclid Ave., St. Louis, MO, 63110 , United States of America, 4 Lifestyle Intervention Research Core, Institute for Clinical and Translational Sciences, Washington University School of Medicine , 660 South Euclid Ave., St. Louis, MO, 63110 , United States of America
1 Academic Editor: Stephen L. Clarke, Oklahoma State University , UNITED STATES
Dietary phytosterols, plant sterols structurally similar to cholesterol, reduce intestinal
cholesterol absorption and have many other potentially beneficial biological effects in humans. Due
to limited information on phytosterol levels in foods, however, it is difficult to quantify habitual
dietary phytosterol intake (DPI). Therefore, we sought to identify a plasma biomarker of DPI.
Methods and Findings
Data were analyzed from two feeding studies with a total of 38 subjects during 94 dietary
periods. DPI was carefully controlled at low, intermediate, and high levels. Plasma levels of
phytosterols and cholesterol metabolites were assessed at the end of each diet period.
Based on simple ordinary least squares regression analysis, the best biomarker for DPI
greatly among subjects at the same DPI level, but were positively correlated at each DPI
level in both studies (r > 0.600; P < 0.01).
The ratio of plasma campesterol to the coordinately regulated endogenous cholesterol
mephytosterol levels alone are not ideal biomarkers of DPI because they are confounded by
large inter-individual variation in absorption and turnover of non-cholesterol sterols. Further
work is needed to assess the relation between non-cholesterol sterol metabolism and
associated cholesterol transport in the genesis of coronary heart disease.
Competing Interests: Washington University and Dr.
Ostlund have a financial interest in Lifeline
Technologies, Inc., a startup company
commercializing emulsified phytosterols. Emulsified
phytosterols and Lifeline products were not used in
this work. This does not alter the authors’ adherence
to PLOS ONE policies on sharing data and materials.
In addition to their robust effects on cholesterol metabolism [1–3], phytosterols are reported to
have many other biological actions, such as anti-inflammatory and anti-cancer effects [4,5].
However, it is difficult to quantify habitual dietary phytosterol intake (DPI) because phytosterol levels in
foods are not systematically included in food databases. Even though food frequency questionnaire
has been used to recall food intake of nutrients, it is more subjective in nature and more
challenging, especially for older populations . Therefore, a suitable plasma biomarker for DPI would be
helpful in assessing the biological effects of dietary phytosterols in epidemiological studies.
Many diet-related investigations rely on measurement of plasma levels of the nutrient being
studied (phytosterols in this case) in order to estimate habitual dietary intake. At first glance,
that approach seems attractive with respect to phytosterols because they are not synthesized in
humans. Therefore any phytosterols measured in plasma must have originated in the diet.
However, phytosterols differ from other nutrients in a fundamental way; their plasma levels
reflect a very complex relationship not only to dietary intake but also to several metabolic
variables. Plasma phytosterol levels are affected by their intestinal absorption efficiency, mediated
by intestinal lipid transporters Niemann-Pick C1-Like 1 (NPC1L1) (favoring sterol uptake) 
and ATP-binding cassette G5 and G8 (ABCG5/ABCG8) transporters (favoring phytosterol
efflux from enterocytes back to the lumen) [8,9]. Plasma phytosterols also are affected by hepatic
NPC1L1 and ABCG5/ABCG8 transporters , which result in the rapid and near complete
biliary excretion of phytosterols. Moreover, increasing phytosterol intake results in a plateau in
plasma phytosterol levels, indicating competition of phytosterols for their own absorption
[11,12]. Taken together, these results suggest that plasma phytosterol levels are dependent not
only on DPI, but also on the absorption efficiency and the re-excretion rate of absorbed
phytosterols. Therefore, plasma phytosterol levels alone are not ideal biomarkers for DPI.
The endogenous cholesterol metabolite 5-α-cholestanol (5α-cholestan-3β-ol), an
intermediate of bile acid synthesis, derives primarily from the catabolism of cholesterol in the liver
[13,14]. Dietary intake of 5-α-cholestanol is very low and its contribution to plasma levels is
thought to be negligible . Phytosterols (e.g., campesterol, sitosterol) and 5-α-cholestanol
are non-cholesterol sterols that are structurally similar to cholesterol. The major differences are
saturation of the Δ5 double bond at the 5α position (5-α-cholestanol) and methyl and ethyl
groups at position C-24 (campesterol and sitosterol, respectively) (Fig. 1). Average cholesterol
absorption efficiency in humans is 56%, whereas phytosterols are not absorbed to a great
extent, with absorption efficiency of campesterol being only 1.9% . Campesterol is a common
dietary phytosterol and has the highest intestinal absorption efficiency among individual
phytosterols. In comparison, the intestinal absorption efficiency of the metabolite 5-α-cholestanol
is 3.3% , resembling that of phytosterols rather than cholesterol.
The aim of this study was to identify a plasma biomarker for dietary phytosterol intake. We
hypothesized that plasma phytosterol levels (total phytosterols or campesterol) would be
suitable after normalization by a marker that reflects overall handling (i.e., absorption and
excretion) of non-cholesterol sterols by the body.
The current study is an analysis of results of two independent clinical studies in which we
investigated the effects on cholesterol metabolism from supplemental phytosterols (Supplement
Study, NCT00860054)  and phytosterols naturally present in the diet (Natural Study,
NCT00860509) . Participant characteristics, study designs, and outcome measures have
been published [2,3]. Briefly, both studies were randomized, cross-over feeding trials in which
all meals were prepared in a metabolic kitchen.
Fig 1. Sterol structures. Top panel shows the structures of cholesterol and the phytosterols campesterol
and sitosterol. Bottom panel shows the structure of 5-α-cholestanol, the reduced form of cholesterol.
In the Supplement Study, 18 adults received a low-phytosterol diet (50 mg phytosterols/
2000 kcal) plus beverages supplemented with 0, 400, or 2000 mg phytosterols/day for 4 weeks
each, in random order. In the Natural Study, 20 subjects consumed two diets for 4 weeks each.
The diets differed in phytosterol content from natural foods (126 mg phytosterols/2000 kcal vs.
449 mg phytosterols/2000 kcal), but otherwise were matched for nutrient content. In both
studies, concentrations of plasma phytosterols, 5-α-cholestanol, and the cholesterol precursor
lathosterol were determined by gas chromatography—mass spectrometry [2,3].
Statistical analysis was performed with SAS software (Version 9.3, SAS Institute Inc., Cary,
NC) on data from the two feeding studies [2,3]. For each study, one-way ANOVA in the Mixed
procedure was used to analyze treatment effects on the observed dietary phytosterol intake
Values are means ± SD.
with Tukey adjustment for multiple comparisons. Spearman’s rank-order correlation was
performed between plasma 5-α-cholestanol and campesterol levels at each DPI level, and between
plasma 5-α-cholestanol and DPI in each study. Linear association of campesterol on
5-α-cholestanol was analyzed with mixed random effects repeated measures regression with a
5-α-cholestanol by treatment interaction term to test differences in slopes between treatments.
Combined data from the 2 studies were analyzed using repeated-measures regression models.
R-Square statistics of simple ordinary least squares regression were used to compare outcome
variables for strength of relationship to DPI with DPI as the independent variable. Dependent
variables included individual phytosterols, total phytosterols, and their ratios to cholesterol and
to 5-α-cholestanol. Repeated measures regression models were fit with the Mixed Procedure
and included period nested within study as the repeated effect and a first order autoregressive
correlation structure within subject. Data were log transformed when appropriate.
Model : LNðYijÞ ¼ b0 þ b1½LNðDPIÞ þ eij;
where the dependent variable LN(Yij) is the natural log of Yij, which represents the ratio of
plasma total phytosterol levels or campesterol levels to 5-α-cholestanol, β0 is the intercept, β1 is the
slope, LN(DPI) is the natural log of DPI, and eij is the residual of subject i in period j.
Subject characteristics at baseline are presented in Table 1. Based on the design of the
controlled feeding studies and excellent adherence to the provided diets, the observed DPI differed
between diet conditions. In the Supplement Study, DPI values at the three phytosterol doses (i.
e., 50, 450, and 2050 mg/2000 Kcal) averaged 99 ± 82 (SEM), 520 ± 82, and 2244 ± 82 mg/day,
respectively (Fig. 2A). In the Natural Study, DPI at the two phytosterol doses (i.e., 126 and 449
mg/2000 Kcal) averaged 101 ± 18 and 632 ± 18 mg/day (Fig. 2B).
At each level of DPI, there was wide variability of plasma campesterol (1.26–16.49,
3.80–18.07, and 5.62–23.17 μg/mL at the three phytosterol doses in the Supplement Study;
1.38–6.48 and 2.88–11.14 μg/mL at the two phytosterol doses in the Natural Study) and
5-αcholestanol concentrations (1.22–6.81, 1.12–4.19, 0.96–3.14 μg/mL, respectively, for
Fig 2. Observed dietary phytosterol intake (DPI) (A, Supplement Study; B, Natural Study). *significantly
different from 50 mg/2000 Kcal, P < 0.01; **significantly different from 50 mg/2000 Kcal or 450 mg/2000 Kcal
(A), or from 126 mg/2000 Kcal (B), P < 0.0001.
Supplement Study; 1.20–4.45 and 0.73–2.55 μg/mL for Natural Study), suggesting that
individual subjects handle non-cholesterol sterols differently (Fig. 3). In addition, plasma campesterol
and 5-α-cholestanol were positively correlated at each of the three DPI levels in the
Supplement Study (Fig. 3A), and at both DPI levels in the Natural Study (Fig. 3B). In the overall data,
there also was a significant and positive association between plasma campesterol and plasma
5-α-cholestanol (P < 0.001) and the slope of this relation differed by phytosterol dose (P <
0.0001). On the other hand, there was a significant negative correlation between plasma
5α-cholestanol and DPI in the Supplement Study (r = -0.4240, P = 0.0014) and in the Natural
Study (r = -0.5947, P < 0.0001).
As shown in Table 2, the variable most strongly correlated with DPI was the ratio of plasma
campesterol to 5-α-cholestanol (R2 = 0.785, P < 0.0001), followed by the ratio of plasma total
phytosterols to 5-α-cholestanol (R2 = 0.767, P < 0.0001). Each of the variables, when normalized by
cholesterol, showed weaker correlations with DPI. The R2 was lowest for each of the non-cholesterol
sterols alone. Repeated measures regression of plasma campesterol/5-α-cholestanol (Fig. 4A, P <
0.0001) and total phytosterols/5-α-cholestanol (Fig. 4B, P < 0.0001) on DPI were highly significant.
Dietary phytosterol intake cannot be quantified accurately in most studies because the phytosterol
content of many individual foods is not available in major food databases. Therefore, we sought to
identify a plasma biomarker for DPI. The principal finding is that the most suitable biomarker of
DPI is the ratio of plasma campesterol to the cholesterol metabolite 5-α-cholestanol. The ratio of
total plasma phytosterols to 5-α-cholestanol also was highly correlated with DPI.
Plasma phytosterol levels alone are not ideal biomarkers for DPI because circulating
phytosterols reflect dietary intake, sterol absorption efficiency, and biliary excretion rate of absorbed
phytosterols, the latter two of which comprise overall phytosterol handling by the body.
Moreover, dietary phytosterols reduce the absorption efficiency of cholesterol, and probably the
absorption efficiency of phytosterols themselves and other non-cholesterol sterols such as
5-αcholestanol. At high levels of DPI, plasma phytosterol levels increase but this rise is blunted due
to lower sterol absorption efficiency from a higher DPI or due to a lower intrinsic absorption
capacity, which underestimates DPI. Conversely, higher sterol absorption efficiency, resulting from
Fig 3. Correlations between plasma non-cholesterol sterol concentrations at each dietary phytosterol
intake level. Values reflect Spearman’s rank-order correlation coefficients in the Supplement Study (A) and
the Natural Study (B).
low DPI or a higher intrinsic absorption capacity, raises plasma phytosterol levels further, which
overestimates DPI. Therefore, normalizing plasma phytosterols by a marker of non-cholesterol
sterol metabolism is expected to improve the estimation of DPI. In our study, the coefficient of
determination (R2) for DPI using ordinary least squares regression analysis increased when
plasma campesterol and total phytosterols were normalized by plasma 5-α-cholestanol (Table 2).
It has been recognized that large variability exists in the response of serum cholesterol levels
to dietary modification or drug therapy among individuals . Similarly, large
inter-individual variability exists in the handling of non-cholesterol sterols, as demonstrated by varied plasma
campesterol concentrations at the same DPI level (expressed as mg phytosterols per kg of body
weight) (Fig. 3). Interestingly, plasma 5-α-cholestanol levels also varied greatly at the same DPI
level (Fig. 3). More importantly, plasma campesterol and 5-α-cholestanol were positively
correlated at each phytosterol dose in both studies, suggesting that 5-α-cholestanol is handled
similarly to phytosterols by lipid transporters. This same positive association also was found in the
combined data in our studies and previously in free-living middle-aged men where DPI was
not controlled . Coordinate regulation of phytosterols and 5-α-cholestanol also has been
reported in animals with ABCG5/8 deficiency, where both phytosterols and 5-α-cholestanol
were increased in plasma and tissues, whereas cholesterol was reduced .
It is interesting to note that plasma 5-α-cholestanol was negatively associated with DPI in
this study, consistent with a previous negative correlation between plasma 5-α-cholestanol and
fecal phytosterols (the latter reflecting dietary phytosterol intake) in middle-aged men .
This strengthened the notion that plasma campesterol and 5-α-cholestanol are indeed
coordinately regulated and this regulation occurred within a wide range of dietary phytosterol intakes.
Plasma 5-α-cholestanol and phytosterols are regulated by similar factors in absorption and
excretion except that 5α-cholestanol is synthesized by the body but phytosterols are not . In
addition, phytosterols appear to compete with the absorption of 5-α-cholestanol as suggested
by a previous positive correlation between dietary phytosterols and fecal 5-α-cholestanol .
The diverging slopes between different DPI levels in Fig. 3 indicate that plasma 5-α-cholestanol
Fig 4. Regression of plasma phytosterol levels normalized by 5-α-cholestanol on dietary phytosterol
intake (DPI). All values were transformed to natural log. Dotted lines represent 95% prediction limits of Ln
(Campesterol/5-α-cholestanol) (A) and Ln(Total phytosterols/5-α-cholestanol) (B).
level rose at a slower rate at a higher DPI. The differences in slope may be related to
competihas been used as a marker for cholesterol absorption efficiency in healthy individuals, as well as
in those with hyperlipidemia and type 2 diabetes [7,9,15,20,21]. In our study, however, the
positive correlation of 5-α-cholestanol with % cholesterol absorption was relatively small by
itself and when normalized by cholesterol (data not shown). Furthermore, there was no
correlation between % cholesterol absorption and plasma 5-α-cholestanol when data were analyzed
by DPI level (data not shown). Finally, the R2 was smaller (0.546) when plasma campesterol
was normalized by the measured cholesterol absorption rather than plasma 5-α-cholestanol
(0.785). Therefore, plasma 5-α-cholestanol appears to be more closely related to overall
handing of phytosterols, rather than cholesterol, by the body.
The data from our controlled feeding studies enabled us to identify a plasma biomarker for
DPI. These results need to be extended to a variety of diets that are typically consumed and in
larger samples. For diets with different phytosterol compositions, the ratio of plasma total
phytosterols to 5α-cholestanol might be preferred. Determination of non-cholesterol sterols must
be performed by mass spectrometry and more work is needed to reduce costs and increase the
standardization of these relatively new methods .
In addition to their robust effects on cholesterol metabolism, phytosterols may have many
other benefits, such as anti-inflammatory and anti-cancer effects [4,5]. The plasma biomarker
identified herein may prove to be useful in investigating the effects of dietary phytosterols on
many biological, physiological, and pathological processes.
Conceived and designed the experiments: REO SBR CAS XL. Performed the experiments: XL
LM. Analyzed the data: MW. Wrote the paper: XL REO SBR MW. Acquisition of data by GC/
1. Demonty I , Ras RT , van der Knaap HCM , Duchateau GSMJE , Meijer L , et al. ( 2009 ) Continuous DoseResponse Relationship of the LDL-Cholesterol-Lowering Effect of Phytosterol Intake . J Nutr 139 : 271 - 284 . doi: 10.3945/jn. 108.095125 PMID: 19091798
2. Racette SB , Lin X , Lefevre M , Spearie CA , Most MM , et al. ( 2010 ) Dose effects of dietary phytosterols on cholesterol metabolism: a controlled feeding study . Am J Clin Nutr 91 : 32 - 38 . doi: 10.3945/ajcn. 2009.28070 PMID: 19889819
3. Lin X , Racette SB , Lefevre M , Spearie CA , Most M , et al. ( 2010 ) The effects of phytosterols present in natural food matrices on cholesterol metabolism and LDL-cholesterol: a controlled feeding trial . European journal of clinical nutrition 64 : 1481 - 1487 . doi: 10.1038/ejcn.2010.180 PMID: 20808333
4. Othman RA , Moghadasian MH ( 2011 ) Beyond cholesterol-lowering effects of plant sterols: clinical and experimental evidence of anti-inflammatory properties . Nutrition reviews 69 : 371 - 382 . doi: 10.1111/j. 1753- 4887 . 2011 .00399.x PMID: 21729090
5. Jones PJ , AbuMweis SS ( 2009 ) Phytosterols as functional food ingredients: linkages to cardiovascular disease and cancer. Current opinion in clinical nutrition and metabolic care 12 : 147 - 151 . PMID: 19209468
6. Adamson AJ , Collerton J , Davies K , Foster E , Jagger C , et al. ( 2009 ) Nutrition in advanced age: dietary assessment in the Newcastle 85+ study . European journal of clinical nutrition 63 Suppl 1 : S6 - 18 . doi: 10.1038/ejcn.2008.60 PMID: 19190647
7. Betters JL , Yu L ( 2010 ) NPC1L1 and cholesterol transport . FEBS letters 584 : 2740 - 2747 . doi: 10. 1016/j.febslet. 2010 . 03.030 PMID: 20307540
8. Kidambi S , Patel SB ( 2008 ) Cholesterol and non-cholesterol sterol transporters: ABCG5, ABCG8 and NPC1L1: a review. Xenobiotica; the fate of foreign compounds in biological systems 38 : 1119 - 1139 . doi: 10.1080/00498250802007930 PMID: 18668442
9. Igel M , Giesa U , Lutjohann D , von Bergmann K ( 2003 ) Comparison of the intestinal uptake of cholesterol, plant sterols, and stanols in mice . Journal of lipid research 44: 533-538. PMID: 12562824
10. Brown JM , Yu L ( 2009 ) Opposing Gatekeepers of Apical Sterol Transport: Niemann-Pick C1-Like 1 (NPC1L1) and ATP-Binding Cassette Transporters G5 and G8 (ABCG5/ABCG8) . Immunology, endocrine & metabolic agents in medicinal chemistry 9: 18 - 29 . doi: 10.5114/pdia.2014.47118 PMID: 25610349
11. Ntanios FY , van de Kooij AJ , de Deckere EA , Duchateau GS , Trautwein EA ( 2003 ) Effects of various amounts of dietary plant sterol esters on plasma and hepatic sterol concentration and aortic foam cell formation of cholesterol-fed hamsters . Atherosclerosis 169 : 41 - 50 . PMID: 12860249
12. Hallikainen M , Simonen P , Gylling H ( 2014 ) Cholesterol metabolism and serum non-cholesterol sterols: summary of 13 plant stanol ester interventions . Lipids in health and disease 13: 72. doi: 10.1186/1476- 511X-13-72 PMID: 24766766
13. Salen G , Shefer S , Tint GS ( 1984 ) Transformation of 4-cholesten-3-one and 7 alpha-hydroxy-4-cholesten-3-one into cholestanol and bile acids in cerebrotendinous xanthomatosis . Gastroenterology 87 : 276 - 283 . PMID: 6735073
14. Jew S , AbuMweis SS , Jones PJ ( 2009 ) Evolution of the human diet: linking our ancestral diet to modern functional foods as a means of chronic disease prevention . Journal of medicinal food 12 : 925 - 934 . doi: 10.1089/jmf.2008.0268 PMID: 19857053
15. Miettinen TA , Tilvis RS , Kesäniemi YA ( 1989 ) Serum cholestanol and plant sterol levels in relation to cholesterol metabolism in middle-aged men . Metabolism 38 : 136 - 140 . PMID: 2913464
16. Ostlund RE Jr, Lin X ( 2006 ) Regulation of cholesterol absorption by phytosterols . Current atherosclerosis reports 8 : 487 - 491 . PMID: 17045075
17. Bhattacharyya AK , Lin DS , Connor WE ( 2007 ) Cholestanol metabolism in patients with cerebrotendinous xanthomatosis: absorption, turnover, and tissue deposition . Journal of lipid research 48: 185-192. PMID: 17012751
18. Schaefer EJ , Lamon-Fava S , Ausman LM , Ordovas JM , Clevidence BA , et al. ( 1997 ) Individual variability in lipoprotein cholesterol response to National Cholesterol Education Program Step 2 diets . The American journal of clinical nutrition 65: 823-830. PMID: 9062535
19. Seyama Y ( 2003 ) Cholestanol metabolism, molecular pathology, and nutritional implications . Journal of medicinal food 6: 217-224. PMID: 14585188
20. Miettinen TA , Tilvis RS , Kesaniemi YA ( 1990 ) Serum plant sterols and cholesterol precursors reflect cholesterol absorption and synthesis in volunteers of a randomly selected male population . American journal of epidemiology 131 : 20 - 31 . PMID: 2293749
21. Tilvis RS , Miettinen TA ( 1986 ) Serum plant sterols and their relation to cholesterol absorption . The American journal of clinical nutrition 43: 92-97. PMID: 3942097
22. Marinangeli CP , Kassis AN , Jones PJ ( 2009 ) Glycemic responses and sensory characteristics of whole yellow pea flour added to novel functional foods . Journal of food science 74 : S385 - 389 . doi: 10.1111/j. 1750- 3841 . 2009 .01347.x PMID: 20492127