Clinica Chimica Acta 414 (2012) 241–245

Contents lists available at SciVerse ScienceDirect

Clinica Chimica Acta

journal homepage: www.elsevier.com/locate/clinchim

Shift of high-density lipoprotein size distribution toward large particles in patients

with proteinuria
Ernesto Soto-Miranda a, Elizabeth Carreón-Torres a, Karina Lorenzo a, Berenice Bazán-Salinas a,
Cynthia García-Sánchez a, Martha Franco b, Carlos Posadas-Romero c, José-Manuel Fragoso a,
Victoria López-Olmos a, Magdalena Madero b, José-Manuel Rodriguez-Pérez a,
Gilberto Vargas-Alarcón a, Oscar Pérez-Méndez a,⁎
a
Department of Molecular Biology, Instituto Nacional de Cardiología “Ignacio Chávez”, Mexico D.F., Mexico
b
Department of Nephrology, Instituto Nacional de Cardiología “Ignacio Chávez”, Mexico D.F., Mexico
c
Department of Endocrinology, Instituto Nacional de Cardiología “Ignacio Chávez”, Mexico D.F., Mexico

a r t i c l e i n f o a b s t r a c t

Article history: Background: The potential atheroprotective role of the different HDL subclasses may depend on the metabolic
Received 21 July 2012 factors that affect their plasma concentrations. The kidney is supposed to be one of the main catabolic sites
Received in revised form 21 September 2012 for these lipoproteins. However, little is known about the impact of proteinuria on HDL size distribution
Accepted 24 September 2012 and HDL structure. The aim of this study is to establish the influence of proteinuria on HDL size distribution
Available online 2 October 2012
and cholesterol plasma concentration of HDL subclasses.
Methods: Forty patients within a range of proteinuria from 0.2 to 10.0 g/g estimated by the urinary
Keywords:
Chronic kidney disease
protein-to-creatinine ratio and 40 healthy controls were enrolled in the study. HDL subclasses were separated
Urinary protein-to-creatinine ratio by sequential ultracentrifugation followed by a polyacrylamide gradient electrophoresis; gels were stained enzy-
Reverse cholesterol transport matically for cholesterol and with Coomasie blue for proteins. HDL size distribution and plasma concentration of
High-density lipoproteins subclasses the five HDL subclasses were calculated by optical densitometry.
Coronary heart disease Results: When determined by protein, large HDL2b and HDL2a relative proportions were higher in patients than
in control subjects, whereas the contrary was observed for small HDL3b and 3c. Consistently, HDL3a, 3b, and 3c
were negatively correlated with proteinuria when data were adjusted by age, gender, body mass index, and
blood pressure. Size distribution followed a different pattern when determined by cholesterol, suggesting an ab-
normal lipid composition that was further supported by a protein-to-cholesterol ratio significantly higher in
most of the HDL subclasses in proteinuric patients than in the control group. Moreover, proteinuria statistically
explains the HDL2b and HDL3c cholesterol plasma concentrations.
Conclusions: Proteinuria is associated with a shift of HDL size distribution towards large particles and
cholesterol-poor HDL subclasses. These results support the idea of a selective loss by the kidney of small HDL
in patients with proteinuria; whether these abnormalities reflect an impaired reverse cholesterol transport
and an increased risk of coronary heart disease remains to be elucidated.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction
Several clinical and epidemiological studies have demonstrated
the negative correlation between HDL-cholesterol (HDL-C) and the
risk of coronary heart disease (CHD). The protective effect of HDL par-
ticles against CHD has been explained by their ability to transport
cholesterol from the peripheral tissues to the liver to be eliminated
Abbreviations: CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration for-
mula; PrU/CrU, urinary protein-to-creatinine ratio.
⁎ Corresponding author at: Department of Molecular Biology, Instituto Nacional de
Cardiología “Ignacio Chávez” Juan Badiano 1, Sección XVI, 14080 Mexico D.F., Mexico.
Tel.: +52 55 55 73 29 11x1278; fax: +52 55 55 73 09 26.
E-mail address: opmendez@yahoo.com (O. Pérez-Méndez).
through the bile. In addition, HDL have antioxidant and anti-
inflammatory actions [1,2].
HDL include a heterogeneous group of lipoproteins that may be
classified by size (in decreasing order) in HDL2b, HDL2a, HDL3a,
HDL3b, HDL3c [3]. These HDL subclasses differ in their physicochem-
ical properties, and it has been suggested that they have different
antiatherogenic characteristics [4,5], the small particles being the
most cardioprotective lipoproteins [5,6].
HDL structure seems to be also determinant for their intravascular
metabolism and removal from the plasma compartment; it is likely
that small HDLs are catabolized faster than larger ones in animal
models [7,8]. Indeed, several studies have suggested a relationship
between HDL structure and apo AI turnover [9–12]. In this context,
the kidney is usually not regarded as an important organ in HDL

0009-8981/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cca.2012.09.028
242 E. Soto-Miranda et al. / Clinica Chimica Acta 414 (2012) 241–245
metabolism because the glomerular filtration barrier prevents large
molecules such as lipoproteins from being filtered into Bowmans's
space. Nevertheless, the proximal tubule epithelium expresses several
receptors involved in the uptake of a broad spectrum of ligands, in-
cluding apolipoproteins [13,14]. Furthermore, impaired renal function
is associated with abnormal HDL-cholesterol and apo AI plasma con-
centrations [15–17]. These changes could be the result of an increased
glomerular permeability that enhances the excretion rate of HDL. In
that case, it could be postulated that small HDL particles are suscepti-
ble to be easily filtrated and further eliminated by the tubules, increas-
ing the CHD risk, whereas large HDL may accumulate in the plasma.
However, little is known about HDL subclasses in proteinuric patients.
Therefore, the aim of this study was to analyze HDL size distribution
and plasma cholesterol concentration of HDL subclasses in patients
within a wide range of proteinuria to understand better the relation-
ship between renal function and lipoprotein profile. Our results support
the idea of a selective loss by the kidney of small HDL in patients with
proteinuria.
2. Methods
2.1. Study Subjects
Patients aged between 18 and 60 years, with chronic kidney disease
(CKD) according to KDOQI classification [18], estimated glomerular fil-
tration rate (eGFR) by CKD-EPI (Chronic Kidney Disease Epidemiology
Collaboration) formula less than 60 ml/min/1.73 m 2 [19], and protein-
uria estimated by the urinary protein/urinary creatinine ratio (PrU/
CrU)> 0.2 g/g were recruited in the Department of Nephrology of the
National Institute of Cardiology, Mexico. The main criterion to select pa-
tients and controls was the protein-to-creatinine ratio, which has been
recognized to be an accurate estimation of proteinuria, sometimes even
better than measuring protein in a 24-h urine sample because it is not
biased by collection deficiencies, mainly attributable to patients. This
is particularly important for external ambulant subjects who are some-
times unable to collect correctly their samples along the day [20]. We
excluded patients who met diabetes criteria, were under current treat-
ment with steroids, statins, fibrates, niacin or those who refused to par-
ticipate in the study. Forty volunteers without personal or family history
of diabetes, pancreatitis, hypertension, angina pectoris, CHD, and who
were not under any pharmacological treatment at least during the
least 4 weeks before the study, were part of the control group. Control
subjects were required also to have a fasting glucose of b110 mg/dl,
BMI of b 32 kg/m2, total cholesterol of b 200 mg/dl, triglycerides of
b200 mg/dl, smoked b 5 cigarettes/day, a calcium score = 0 in Agatston
score determined by computed tomography [21], as well as normal he-
patic and thyroid functions as assessed by routine laboratory analyses,
and kidney function (eGFR>60 ml/min/1.73 m2, and PrU/CrUb 0.1 g/g).
Control subjects were recruited from an institutional study of Genetics
of Cardiovascular Diseases during 2009 and 2010.
All included individuals were asked to avoid extreme physical activ-
ity the day before the analysis. Physical examination was performed to
each patient and included measurement of vital signs, BMI, and cardio-
vascular examination. In addition, 24-h urine collection was performed
on each patient at the same time of blood drawing. Random urine sam-
ples (first urine sample in the morning) were collected from both con-
trol subjects and patients. The study was approved by the Ethics
Committee of the National Institute of Cardiology of Mexico. All eligi-
ble participants for the study were informed of its objectives and
those who agreed to participate signed a letter of informed consent.
2.2. Laboratory assessment
All patients were instructed to avoid strenuous exercise and to eat
a light dinner the day before blood drawings were performed. Blood
samples were obtained in EDTA tubes after 12-h overnight fasting,
centrifuged at 4 °C, the plasma was separated and analyzed or frozen
at − 80 °C until analysis. Plasma uric acid, glucose, total cholesterol,
and triglycerides (Tg) were determined by commercially enzymatic
available colorimetric methods (Randox LTD, Crumlin, UK). The
phosphotungstic acid-Mg2+ method was used to precipitate apoB-
containing lipoproteins before quantifying HDL-cholesterol plasma con-
centrations (Randox LTD). Low-density lipoprotein-cholesterol (LDL-C)
was estimated for samples with triglycerides b 400 mg/dl [22]. Plasma
concentrations of all lipids were determined within 24 h after drawing
blood samples. Urine protein was determined in the 24-h urine sample
using the trichloroacetic acid assay [23]. Urine and plasma creatinine
were measured by the modified kinetic Jaffe reaction (Randox LTD).
2.3. Isolation of HDL and enzymatic cholesterol staining on
polyacrylamide gels
HDL were separated by ultracentrifugation in a Beckman optima
TLX table centrifuge (Beckman Instruments Inc., Palo Alto, CA) at
110,000 rpm in 3.2 ml polycarbonate tubes as described previously [7].
Briefly, total apo B-containing lipoproteins (densityb 1.063 mg/dl) were
obtained after 2.16 h, whereas total HDL (1.063b densityb 1.21 g/ml)
took 2.5 h. Under these conditions, 80 to 85% of total plasma apo A-I
was recovered from the HDL fraction without apo B-contamination.
HDL were dialyzed against 0.09 M Tris/0.08 M boric acid/3 mM EDTA
buffer, pH 8.4.
HDL were further separated by their hydrodynamic diameter in a
non-denaturing 3–30% gradient polyacrylamide gel electrophoresis,
as previously described [7,24,25]. Twenty-five micrograms of HDL
protein sample, corresponding approximately to 10 μg of cholesterol,
was deposited per well.
Gels were stained for cholesterol using an enzymatic mixture in
carboxymethylcellulose at 1.4%, as recently described [25]. Gels were
stained for cholesterol using an enzymatic mixture of cholesterol es-
terase, cholesterol oxidase, and peroxidase at a final concentration of
0.075 U/ml, 0.05 U/ml, and 0.25 U/ml, respectively, in a 150 mM
NaCl, 8.6 mM Na2HPO4, 1.4 mM NaH2PO4 buffer (PBS), pH 7.4. The re-
action mixture also included 3 mM sodium cholate, 0.1% Triton 100×,
0.4 mM thiazolyl blue teratozolium bromide, 0.6 mM phenazine
methosulphate, and carboxymethylcellulose at 1.4% (Sigma-Aldrich,
St. Louis, MO). Electrophoresis gels were incubated with the reaction
mixture during 1 h at 37 °C in the dark, and gently washed in PBS at
the end of the incubation time. Electrophoresis gels were then scanned
in a GS-670 BioRad densitometer (BioRad Laboratories Inc., Hercules,
CA) (scan 1), distained, further re-stained for proteins with Coomassie
R-250, and scanned again (scan 2). The relative proportions of each HDL
subclass determined by protein were estimated by optical densitometry
analysis of scan 2, used as reference globular proteins (thyroglobulin,
17 nm; ferritin, 12.2 nm; catalase 10.4 nm; lactate dehydrogenase,
8.2 nm; and albumin, 7.1 nm; high-molecular weight calibration kit,
Amersham Pharmacia Biotech, Buckimghamshire, UK). Relative pro-
portion of each HDL subclass is expressed as the percentage of the
total HDL area under the curve, integrated from 7.94 to 13.59 nm.
The relative proportion of HDL subclasses according to their choles-
terol content was determined in scan 1 using the migration distances of
the reference globular proteins obtained from scan 2. Considering that
the area under the curve in the densitogram represents 100% of the cho-
lesterol in the HDL, the cholesterol plasma concentration of each HDL
subclass was estimated as follows:
HDLn−C ¼ ð% HDLn determined by cholesterol  HDL−CÞ=100
where n represents the HDL subclass, and HDL-C is the HDL-cholesterol
plasma concentration.
For classification of the HDL subclasses, we considered the following
size intervals: HDL3c, 7.94–8.45 nm; HDL3b, 8.45–8.98 nm; HDL3a,
 E. Soto-Miranda et al. / Clinica Chimica Acta 414 (2012) 241–245 243

8.98–9.94 nm; HDL2a, 9.94–10.58 nm; and HDL2b, 10.58–13.59 nm 0.36 ± 0.09 (P b 0.001) in the control group, suggesting structural
[7,24]. changes of these lipoproteins.
There were no smokers in the group of proteinuric patients,
whereas seven current smokers were included in the control group.
2.4. Statistical analysis
Since smoking may alter lipid profile, blood pressure, and proteinuria,
we performed a sub-analysis excluding current smokers (Supplemen-
Central tendency and dispersion measurements were estimated
tary Table 1, available online). Anthropometric and biochemical dif-
by conventional methods. Comparisons between multiple groups
ferences between groups in this sub-analysis were essentially the
were assessed by ANOVA test. Normal distribution of the variables
same as compared with those presented in Table 1.
was evaluated by the Kolmogorov–Smirnoff test. The significance of
In order to gain more insight about the possible structural differ-
the differences between groups was tested by Student's t test for nor-
ences of HDL in patients with proteinuria, we further analyzed HDL
mally distributed variables. Variables without normal distribution
size distribution determined by protein and by cholesterol; results
were logarithmically transformed for parametric statistical analysis.
are shown in Table 2. When determined by protein staining, we ob-
Comparisons of non-normally distributed variables were performed
served an increment of the relative proportion of large HDL concom-
by Mann–Whitney U for independent groups. Partial correlations
itant with a decreased proportion of small particles. In contrast, the
controlled by age, gender, BMI, systolic and diastolic blood pressure
pattern of size distribution determined by cholesterol content was
were performed and statistical significance was set at P b 0.05. Unless
not clearly shifted towards large particles; the percentage of choles-
otherwise indicated, values are expressed as mean ± SD for variables
terol corresponding to HDL2b was higher in the proteinuric group
with normal distribution and as median and interquartile interval for
than in controls, whereas HDL3a and HDL3b were lower in the study
non-normally distributed variables. Statistical analysis was performed
group (Table 2). Since the different HDL size distribution further suggests
using SPSS v.11 software.
an abnormal HDL structure, we calculated the protein-to-cholesterol
ratio as indicative marker of HDL composition. Through this approach,
3. Results HDL subclasses from proteinuric patients had abnormal cholesterol con-
tent (Table 2); large particles were enriched with cholesterol, whereas
Anthropometric and biochemical data of the subjects included in small HDL3c were cholesterol-poor when compared to the control
the study are presented in Table 1. Mean age and body mass index group. This abnormal composition was accompanied by significant de-
were significantly higher in the control group than in the proteinuric creases only in HDL3a- and HDL3b-cholesterol plasma concentrations
patients. As expected, statistical differences were also found in uric in proteinuric patients while cholesterol concentrations of the
acid plasma concentration, eGFR, and proteinuria whereas no differ- other subclasses remained comparable between the 2 groups (Fig. 1).
ences in blood pressure or glucose were observed; extreme proteinuria In order to discard an effect of smoking on HDL subclasses, we
values were 0.54 and 9.90 g/g in patients, whereas they were 0.05 and performed a sub-analysis including only the 33 non-smoker controls
0.11 g/g for control subjects. Essential hypertension was the most fre- (Supplementary Table 2, available online). Our results clearly demon-
quent diagnosis in proteinuric patients (25%), followed by lupus glomer- strated that the observed differences between proteinuric patients
ulonephritis (17.5%), focal segmental glomerulosclerosis secondary to and non-smoking control subjects (n= 33) remained essentially the
congenital interatrial communication (12.5%), and membranous glo- same as compared to those observed with the whole group of controls
merulonephritis (12.5%). Other diagnoses included post-infectious glo- (n=40).
merulonephritis and amyloidosis. Protein-to-cholesterol ratios of HDL subclasses were not statistically
Cholesterol, HDL-cholesterol, LDL-cholesterol, and triglycerides associated with proteinuria, estimated by PrU/CrU ratio (Spearman's
were similar between groups. In contrast, HDL-triglycerides and correlation P > 0.05 for the five subclasses). In contrast, the relative pro-
HDL-phospholipids were lower in patients as compared to controls portion of large HDL subclasses, determined by their protein content,
(21.7 ± 9.9 mg/dl vs. 36.0 ± 17.5 mg/dl, and 72.0 ± 26.5 mg/dl vs. positively correlated with proteinuria estimated by the ProU/CrU ratio,
131.4 ± 31.7 mg/dl respectively, Pb 0.001 for both). Plasma cholesterol- whereas small HDL3b and 3c presented a negative correlation with
to-phospholipid HDL ratio for the study group was 0.69 ± 0.31 vs. this parameter when controls and patients were analyzed (Table 3).
When the correlation was controlled by non-matched variables be-
Table 1 tween groups (BMI, age, gender, SBP, and DBP), large HDL 2b and 2b
Demographic, anthropometric, and biochemical characteristics of the individuals in- were no longer associated with PrU/CrU, whereas small HDL 3b and
cluded in the study. 3c remained significantly correlated. Interestingly, HDL3a, which was
Study group Control group Pa
not correlated before adjustment by non-matched variables, became
significantly associated with PrU/CrU. The unadjusted correlations had
n= 40 n = 40
a similar trend within each group when controls and patients were an-
Age (y) 35.1 ± 13.0 57.0 ± 8.7 0.000 alyzed independently (Table 3), but only the percentage of protein
Female gender 17 19 NSb
HDL3c in patients and HDL3b in the control group remained statistically
BMI (kg/m2) 25.2 ± 4.8 27.5 ± 3.9 0.031
SBP (mm Hg) 123.8 ± 22.0 123.1 ± 19.0 NS significant. In the group of patients, when data were controlled by the
DBP (mm Hg) 79.6 ± 13.1 74.2 ± 8.2 NS above-mentioned variables, the relative proportion of small HDL parti-
Glucose (mg/dl) 90.1 ± 19 92.7 ± 13.5 NS cles remained statistically associated with PrU/CrU as observed in the
Uric acid (mg/dl) 7.2 ± 1.6 5.5 ± 1.3 0.000 whole group.
Total cholesterol (mg/dl) 186.9 ± 45.0 195.1 ± 36.0 NS
Triglycerides (mg/dl) 167.6 ± 75.0 175.64 ± 81.1 NS
Interestingly, HDL2b- and HDL3c-cholesterol plasma concentrations
LDL-cholesterol (mg/dl) 111.9 ± 32.8 113.5 ± 24.4 NS correlated negatively with proteinuria determined by the PrU/CrU ratio
HDL-cholesterol (mg/dl) 46.2 ± 16.6 48.6 ± 12.6 NS (r= −0.491, and r= −0.462, P b 0.05 for both, respectively) within the
eGFR (ml/min/1.73 m2) 50.5 ± 6.9 89.5 ± 3.2 0.000 group of patients. Such correlation remained significant after adjust-
PrU/CrU (g/g) 1.960 [1.178–3.134] 0.032 [0.014–0.060] 0.000
ment for age, gender, and BMI (r= −0.539, and r= −0.490, P b 0.05
Data are shown as mean ± SD or median [interquartile range]. SBP: systolic blood for both). To gain further insight on the association of proteinuria
pressure, DBP: diastolic blood pressure, eGFR: estimated renal filtration rate measured with HDL2b and HDL3c-cholesterol plasma concentrations, we per-
by the equation of Chronic Kidney Disease Epidemiology Collaboration, NS: not
statistically significant.
formed stepwise regression analysis including markers of kidney func-
a
Proteinuric vs. control group Student's t or Mann–Whitney U test. tion, creatinine plasma concentrations and PrU/CrU, age and BMI, as
b
χ2 test. independent variables. HDL subclasses were included as dependent
244 E. Soto-Miranda et al. / Clinica Chimica Acta 414 (2012) 241–245

Table 2
Relative proportions based on the protein or cholesterol content of HDL subclasses, and the corresponding protein-to cholesterol ratio of individuals included in the study.

% of total HDL-protein % of total HDL-cholesterol Protein-to-cholesterol ratio (%/%)

Proteinuric Control Proteinuric Control Proteinuric Control

HDL2b 23.5 ± 5.9⁎ 15.4 ± 5.7 31.1 ± 8.8⁎⁎ 24.0 ± 5.7 0.777 ± 0.177⁎ 0.629 ± 0.157
HDL2a 11.6 ± 1.87⁎ 8.8 ± 1.9 11.4 ± 1.8 11.8 ± 2.2 1.031 ± 0.147⁎⁎ 0.755 ± 0.183
HDL3a 24.6 ± 2.4 24.1 ± 2.7 19.9 ± 1.7⁎⁎ 24.3 ± 3.8 1.243 ± 0.159⁎⁎ 0.990 ± 0.137
HDL3b 15.6 ± 2.5⁎⁎ 20.6 ± 2.9 12.1 ± 3.7⁎⁎ 16.9 ± 3.8 1.335 ± 0.198 1.272 ± 0.184
HDL3c 24.8 ± 5.7⁎⁎ 31.9 ± 7.3 25.4 ± 7.2 22.9 ± 5.8 1.030 ± 0.284⁎⁎ 1.451 ± 0.320
⁎⁎ t test P b 0.001 vs. control group.
⁎ t test P b 0.005 vs. control group.

variables. In this model, only PrU/CrU predicts HDL2b (β = −0.469,
P =0.032) and HDL3c-cholesterol plasma concentrations (β =−0.438,
P =0.047).
4. Discussion
This study demonstrated that HDL size distribution became
shifted toward large HDL particles in patients with proteinuria, and
the extent of proteinuria is inversely related to the relative proportion
of small HDL particles. Moreover, the cholesterol plasma concentra-
tions of HDL2b and HDL3c subclasses were strongly dependent on
proteinuria.
We included patients within a large range of proteinuria searching
for common HDL characteristics when glomerular membrane perme-
ability is impaired. We excluded subjects with diabetic nephropathy
since type 2 diabetes is associated with an increased proportion of
small HDL particles [26,27], a condition that could have biased our data.
Control subjects were older than proteinuric patients and presented
a higher BMI than patients; since proteinuria tends to increase with age,
and HDL-cholesterol to decrease, this difference could be considered as
an advantage more than a weakness of the study. Nevertheless, correc-
tion by age and BMI were considered for statistical analysis since the
lipid profile is affected by both parameters. When data were corrected
(not shown), all statistical analyses remained as presented in Table 1.
HDL size has been inversely associated with the fractional catabolic
rate of these lipoproteins [7,9,12], meaning that the smaller the particle
size, the higher its clearance from the plasma. Furthermore, it has been
suggested that the kidney is one of the most important catabolic sites
for HDL [14,28]. Based on this evidence, we hypothesized that small
HDL particles have more probabilities to cross the glomerular mem-
brane than large ones. As a consequence, small particles should be easily
removed by glomerular filtration when the permeability of the mem-
brane is increased. Our results of HDL size distribution support this
20
Cholesterol plasma concentration (mg/dL)
hypothesis; when determined by protein content, patients with pro-
teinuria had lower relative proportion of small HDL and higher propor-
tions of large particles than controls, suggesting a leakage of the former
rather than the latter. The negative correlation between small HDL3b
and 3c and proteinuria further supports these data. When the correla-
tion analyses were performed for each independent group, we observed
a similar trend in the patients as that observed for the whole group, but
data did not reach statistical significance. Those correlations between
the three subclasses of HDL3 and the PrU/CrU ratio became significant
when data were adjusted by BMI, age, gender, and blood pressure, in
the whole group, as well as in the proteinuric group; in contrast, corre-
lations with large HDL were lost. On the one hand, these observations
support the idea that mainly small HDL particles are lost by glomerular
filtration as a function of the degree of proteinuria. On the other hand,
the inverse correlations observed between large HDL2b and 2a with
the PrU/CrU ratio seem to be related with age, gender, BMI, and blood
pressure, but not with proteinuria.
The shift of HDL size distribution towards large particles was con-
comitant with alterations of their cholesterol content; cholesterol
plasma concentrations of HDL2b and 3c were strongly associated to
proteinuria, emphasizing the role of the kidney in HDL-cholesterol
homeostasis. Moreover, the protein-to-cholesterol ratio within four
of the five HDL subclasses tended to be lower in proteinuric patients
than in controls. On the basis that HDL size is determined by the num-
ber of phospholipids and proteins on their surface [29,30], and con-
sidering that HDL subclasses are defined by their hydrodynamic
diameters, we postulate that for each HDL subclass there is a specific
and constant number of phospholipids and proteins necessary to
maintain the structure and size of the lipoprotein [31,32]. Therefore,
the interpretation of a lower protein-to-cholesterol ratio is that HDL
are cholesterol-poor in patients. In contrast, the protein-to-cholesterol
ratio suggests a cholesterol enrichment of HDL3c. It is likely that the
cholesterol from the tissues reaches small HDL particles, but these
have a limited residence time in the plasma secondary to renal leakage,
thus delaying the flux of cholesterol to larger particles. Apo AI and cho-
lesterol kinetic studies in proteinuric models are needed to demonstrate

18 the increased catabolism of small HDL particles and the impaired flux of
16 cholesterol from small to large HDL.
We recognize an important limitation in the present study; the in-
14
clusion of patients with heterogeneous pathophysiologies, in some
12 cases concurrent with an inflammation process associated to protein-
uria. We excluded diabetic patients who are known to present HDL
10
size abnormalities [26,27], but it could not be discarded that any
8 other pathology, i.e., hypertension or membranous glomerulonephritis,
may be associated to modifications in HDL subclasses. Moreover, HDL
6
size distribution abnormalities have been reported in some inflamma-
4 tory diseases [33,34], suggesting a contribution of inflammation to
HDL abnormalities. Animal models of proteinuria may help to validate
2
the interpretation of the present data.
0 In summary, in this study we demonstrated an increased relative
HDL2b HDL2a HDL3a HDL3b HDL3c proportion of large HDL and low percentage of small particles in pa-
Fig. 1. Cholesterol plasma concentrations corresponding to the five HDL subclasses in
tients with proteinuria; taken together, these results and the correlation
the proteinuric patients (solid bars) and control subjects (white bars). See Section 2.3 analysis suggest a preferential loss of small HDL in patients with pro-
for details. * t test P b 0.05. teinuria. Moreover, the differences observed in HDL size distribution
 E. Soto-Miranda et al. / Clinica Chimica Acta 414 (2012) 241–245 245

Table 3
Correlation analysis between % of HDL subclasses determined by protein content and PrU/CrU ratio.

All subjects Proteinuric patients Control subjects

(n = 80) (n = 40) (n = 40)

Uncorrecteda Correctedb Uncorrecteda Correctedb Uncorrecteda Correctedb

r/P r/P r/P r/P r/P r/P

HDL 2b protein (%) 0.595/0.000 −0.171/0.251 0.368/0.092 −0.321/0.193 0.310/0.102 0.413/0.045

HDL 2a protein (%) 0.632/0.000 −0.250/0.090 0.395/0.069 −0.428/0.076 0.316/0.095 0.285/0.176
HDL 3a protein (%) −0.009/0.949 −0.501/0.000 0.261/0.240 −0.563/0.015 −0.245/0.201 −0.272/0.199
HDL 3b protein (%) −0.690/0.000 −0.568/0.000 −0.232/0.299 −0.562/0.015 −0.471/0.010 −0.456/0.025
HDL 3c protein (%) −0.466/0.001 −0.468/0.001 −0.453/0.034 −0.636/0.005 −0.040/0.837 −0.095/0.659
a
Spearman's correlation coefficients (r) and statistical significance (P) of data.
b
Correlation analysis controlled by gender, body mass index, systolic blood pressure, diastolic blood pressure, and age.

determined by protein as compared with the distribution estimated by
cholesterol indicated an abnormal lipid composition of HDL in protein-
uric patients. Whether these HDL abnormalities contribute to inefficient
reverse cholesterol transport and to the development of cardiovascular
complications in proteinuric patients remains to be elucidated.
Conclusion
Independent from the etiology of the proteinuria, abnormal renal
loss of protein is associated with a shift of HDL size distribution to-
wards large particles and cholesterol-poor HDL subclasses.
Acknowledgment
This study was supported by a CONACYT grant, number 132473.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.cca.2012.09.028.
References
[1] Navab M, Reddy ST, Van Lenten BJ, Fogelman AM. HDL and cardiovascular disease:
atherogenic and atheroprotective mechanisms. Nat Rev Cardiol 2011;8:222-32.
[2] Rye KA, Barter PJ. Predictive value of different HDL particles for the protection against
or risk of coronary heart disease. Biochim Biophys Acta 2012;1821:473-80.
[3] Williams PT, Krauss RM, Nichols AV, Vranizan KM, Wood PD. Identifying the pre-
dominant peak diameter of high-density and low-density lipoproteins by electro-
phoresis. J Lipid Res 1990;31:1131-9.
[4] Deakin S, Leviev I, Gomaraschi M, Calabresi L, Franceschini G, James RW. Enzy-
matically active paraoxonase-1 is located at the external membrane of producing
cells and released by a high affinity, saturable, desorption mechanism. J Biol Chem
2002;277:4301-8.
[5] Kontush A, Chantepie S, Chapman MJ. Small, dense HDL particles exert potent
protection of atherogenic LDL against oxidative stress. Arterioscler Thromb Vasc
Biol 2003;23:1881-8.
[6] Castro GR, Fielding CJ. Early incorporation of cell-derived cholesterol into
pre-beta-migrating high density lipoprotein. Biochemistry 1988;12(27):25-9.
[7] Huesca-Gómez C, Franco M, Luc G, et al. Chronic hypothyroidism induces abnor-
mal structure of high-density lipoproteins and impaired kinetics of apolipopro-
tein A-I in the rat. Metabolism 2002;51:443-50.
[8] Lee JY, Lanningham-Foster L, Boudyguina EY, et al. Prebeta high density lipoprotein
has two metabolic fates in human apolipoprotein A-I transgenic mice. J Lipid Res
2004;45:716-28.
[9] Carreón-Torres E, Juárez-Meavepeña M, Cardoso-Saldaña G, et al. Pioglitazone in-
creases the fractional catabolic and production rates of high-density lipoproteins
apo AI in the New Zealand white rabbit. Atherosclerosis 2005;181:233-40.
[10] Carreón-Torres E, Rendón-Sauer K, Monter-Garrido M, et al. Rosiglitazone modifies
HDL structure and increases HDL-apo AI synthesis and catabolism. Clin Chim Acta
2009;401:37-41.
[11] Pérez-Méndez O, Álvarez-Salcedo P, Carreón Torres E, et al. Palmitic acid reduces
the fractional catabolic rates of HDL apo A-I in vivo: contribution of the surface
negative charge. Clin Chim Acta 2007;378:53-8.
[12] Pérez-Méndez O, Bruckert E, Franceschini G, et al. Metabolism of apolipoproteins
A-I and A-II in subjects carrying similar apo A-I mutations, apo A-IMilano and apo
A-IParis. Atherosclerosis 2000;148:317-26.
[13] Moestrup SK, Nielsen LB. The role of the kidney in lipid metabolism. Curr Opin
Lipidol 2005;16:301-6.
[14] Glass CK, Pittman RC, Keller GA, Steinberg D. Tissue sites of degradation of
apoprotein A-I in the rat. J Biol Chem 1983;258:7161-7.
[15] Holzer M, Birner-Gruenberger R, Stojakovic T, et al. Uremia alters HDL composi-
tion and function. J Am Soc Nephrol 2011;22:1631-41.
[16] Cho KH, Park SH, Park JE, et al. The function, composition, and particle size of
high-density lipoprotein were severely impaired in an oliguric phase of hemor-
rhagic fever with renal syndrome patients. Clin Biochem 2008;41:56-64.
[17] Vaziri ND. Dyslipidemia of chronic renal failure: the nature, mechanisms, and
potential consequences. Am J Physiol Renal Physiol 2006;290:F262-72.
[18] National Kidney Foundation. NKF K/DOQI guidelines: clinical practice guidelines and
clinical practice recommendations. Available at http://www.kidney.org/professionals/
Kdoqi/guideline_upHD_PD_VA/va_guide1.htm; 2006. (Accessed June 27, 2012).
[19] Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular
filtration rate. Ann Intern Med 2009;150:604-12.
[20] Burden R, Tomson C, Guideline Development Committee, Joint Specialty
Committee on Renal Disease of the Royal College of Physicians of London and
the Renal Association. Identification, management and referral of adults with
chronic kidney disease: concise guidelines. Clin Med 2005;5:635-42.
[21] Agatston AS, Janowitz WR, Hildner FJ, et al. Quantification of coronary artery cal-
cium using ultrafast computed tomography. J Am Coll Cardiol 1990;15:827-32.
[22] DeLong DM, DeLong ER, Wood PD, Lippel K, Rifkind BM. A comparison of methods
for the estimation of plasma low- and very low-density lipoprotein cholesterol.
The Lipid Research Clinics Prevalence Study. JAMA 1986;256:2372-7.
[23] Franco M, Bautista R, Pérez-Méndez O, et al. Renal interstitial adenosine is in-
creased in angiotensin II-induced hypertensive rats. Am J Physiol Renal Physiol
2008;294:F84-92.
[24] Blanche PJ, Gong EL, Forte TM, Nichols AV. Characterization of human
high-density lipoproteins by gradient gel electrophoresis. Biochim Biophys Acta
1981;665:408-19.
[25] Toledo-Ibelles P, García-Sánchez C, Avila-Vazzini N, et al. Enzymatic assessment
of cholesterol on electrophoresis gels for estimating HDL size distribution and
plasma concentrations of HDL subclases. J Lipid Res 2010;51:1610-7.
[26] Garvey WT, Kwon S, Zheng D, et al. Effects of insulin resistance and type 2 diabetes
on lipoprotein subclass particle size and concentration determined by nuclear mag-
netic resonance. Diabetes 2003;52:453-62.
[27] Pérez-Méndez O, Torres-Tamayo M, Posadas-Romero C, et al. Abnormal HDL sub-
classes distribution in overweight children with insulin resistance or type 2 diabetes
mellitus. Clin Chim Acta 2007;376:17-22.
[28] Braschi S, Neville TA, Maugeais C, Ramsamy TA, Seymour R, Sparks DL. Role of the
kidney in regulating the metabolism of HDL in rabbits: evidence that iodination
alters the catabolism of apolipoprotein A-I by the kidney. Biochemistry 2000;39:
5441-9.
[29] Rye KA, Duong MN. Influence of phospholipid depletion on the size, structure, and
remodeling of reconstituted high density lipoproteins. J Lipid Res 2000;41:1640-50.
[30] Gu F, Jones MK, Chen J, et al. Structures of discoidal high density lipoproteins: a
combined computational-experimental approach. J Biol Chem 2010;285(7):
4652-65.
[31] García-Sánchez C, Torres-Tamayo M, Juárez-Meavepeña M, et al. Lipid plasma
concentrations of HDL subclasses determined by enzymatic staining on polyacryl-
amide electrophoresis gels in children with metabolic syndrome. Clin Chim Acta
2011;412:292-8.
[32] Juárez-Meavepeña M, Carreón-Torres E, López-Osorio C, et al. The Srb1 + 1050T
Allele Is Associated with metabolic syndrome in children but not with cholesteryl
ester plasma concentrations of high-density lipoprotein subclasses. Metab Syndr
Relat Disord 2012;10:110-6.
[33] Juárez-Rojas J, Medina-Urrutia A, Posadas-Sánchez R, et al. High-density lipopro-
teins are abnormal in young women with uncomplicated systemic lupus
erythematosus. Lupus 2008;17:981-7.
[34] Espinola-Zavaleta N, Soto-López ME, Carreón-Torres E, et al. Altered flow-mediated
vasodilatation, low paraoxonase-1 activity, and abnormal high-density lipoprotein
subclass distribution in Takayasu's arteritis. Circ J 2009;73:760-6.