CLIN. CHEM. 31/1, 65-69 (1985)
Zinc Content of Cellular Components of Blood: Methods for Cell Separation
and Analysis Evaluated
David B. Mime, Nick V. C. Ralston, and James C. Waliwork
Platelets,
mononucleated
and erythrocytes were
cells, polymorphonucleated
separated
cells,
from whole blood by use of
discontinuous gradients of colloidal polyvinylpyrrolidonecoated silica (“Percoll”). We measured the zinc content of
these cells by flame atomic absorption spectrophotometry,
using a modified technique for micro-samples that obviated
matrix interferences. Thus, results obtained by conventional
flame atomic absorption and by the micro-method were
identical. Inter-comparisons of separation methods indicated
that separation of platelets and mononucleated cells by a
two-gradient system of “FicolI-Hypaque” (a synthetic polymer of sucrose) or Percoll was relatively poor, whereas there
was a good separation when a tertiary gradient system of
Percoll was used. The apparent zinc content of mononucleated cells depended on the degree of separation from the
platelets, with contamination by platelets resulting in artificially high values for mononucleated cells.
Addftlonal Keyphrases: sample preparation
flame atomic absorption spectrophotometiy
reference inteival
The importance
of zinc to human health is well established (1-4). However, present methods for clinical evaluation of zinc status may be inadequate (5). Because concentrations of zinc in plasma or serum-the
most commonly
used measure of zinc status-are subject to a wide variety of
influences that may not necessarily reflect zinc nutriture (5,
6), clinical methods must be developed that more accurately
reflect the status of zinc in the body. It has been suggested
that zinc in blood cells, leukocytes, or erythrocytes may
more accurately reflect tissue zinc (7-9), and procedures for
measuring these have been proposed (8, 10, 11). However,
these methods appear to give inadequate results either
because of poorly defined cell populations or the use of
analytical procedures for zinc that require correction for
strong interference by the matrix.
Here we describe a procedure for separating platelets,
mononucleated cells (MNC), polymorphonucleated
cells
(PMN), and erythrocytes
from a single sample of whole
blood.’ These cells are then analyzed for zinc by flame
atomic absorption spectrophotometry
(FAAS), by a technique that requires microliter quantities of sample and is
not subject to such interference.
Materials and Methods
Apparatus
In all analyses
for zinc we used a Model 5000 atomic
absorption
spectrophotometer
(Perkin-Elmer
Corp., Nor-
‘Nonstandard abbreviations:
MNC, mononucleated
leukocytes;
PMN, polymorphonucleated cells; FAAS, flame atomic absorption
spectrophotometry.
USDA, ARS, Grand Forks Human Nutrition Research Center,
P.O. Box 7166, University Station, Grand Forks, ND 58202.
Received July 2, 1984; accepted October 29, 1984.
walk, CT 06856),2 operated at 213.9 nm in the peak-height
absorbance mode, with the slit width set at 0.7 nm. A singleslot burner head was used, with flame conditions set at the
manufacturer-recommended
settings of 20 for acetylene and
30 for oxidant, which corresponded to flow rates of 1.6 L of
acetylene and 18.0 L of air per minute. The nebulizer was
set to aspirate at 7.0 mL/min.
A Teflon small-volume
sampling cup (Perkin Elmer
Corp., no. 040-7002), attached directly to the nebulizer, was
used throughout all standardization and sample analyses.
Reagents
The water used for solution preparation and sample
was processed through the Millipore
Super Q
(Millipore
Corp., Bedford, MA): an activated charcoal ifiter, two high-capacity demineralizing cartridges, and
a 0.2-sm final ifiter. This water was virtually
ion free
(resistance between 15 and 18 Mfl/cm).
‘Percoll” (Pharmacia Fine Chemicals, Uppsala, Sweden)
is a colloidal solution of polyvinylpyrrolidone-coated
silica
particles, developed specifically as a centrifugation
medium
(12). This colloid is inert, iso-osmotic at different concentrations, and stable at physiological pH and ionic strength.
Fractionated cells and sub-cellular particles maintain their
biological integrity when prepared in gradients of this
material (13).
All chemicals used were of analytical grade (Sigma
Chemical Co., St. Louis, MO). Standards were prepared by
dilution of a 1000 g/g reference solution, a certified atomic
dilution
System
absorption
standard
(J. T. Baker Chemical Co., Phillipsburg, NJ). Acid-washed
glassware was used throughout the
study.
Cell Separation and Sample Preparation
A 9-mL specimen of blood was drawn from fasting subjects into a plastic syringe containing Na2EDTA as anticoagulant (final concentration: 1.5 mg/mL of whole blood).
After removing aliquots for blood-cell counts, we diluted the
remaining sample with an equal volume of isotonic saline,
mixing by careful inversion.
We prepared a discontinuous gradient of Percoll at various concentrations as shown in Figure 1. Three solutions of
different densities of Percoll-1.060, 1.075, and 1.095 kg/L-.were prepared by dilution with isotonic saline. After carefully layering these solutions in a centrifuge tube, heaviest to
lightest, we layered 6 mL of the diluted blood on top of the
gradient and centrifuged for 7 to 10 mm at 1200 x g,
whereupon the cellular components of the blood were separated into four distinct bands (Figure 1). Each layer of cells
was carefl.illy removed and placed in a separate tube.
‘Mention of a trademark or proprietary product does not constitute a guarantee or warranty of the product by USDA and does not
imply its approval to the exclusion of other products that may be
suitable.
CLINICALCHEMISTRY, Vol. 31, No. 1,
1985
65
DENSITY
GRADIENT
suspended in 10 mL of isotonic saline. An aliquot of the
suspended cells was taken for cell counts and hemoglobin
determinations.
Analysis for Zinc
-
C
B
-
-
Whole Blood diluted
1:1 with normal saline
1.060
g/mI
Percoll
1.075
9/mI
Percoll
1.095
g/mI PercoH
AFTER CENTRIFUGATION
10 mm at 1200 X g
Dilute Plasma
Platelet
Cloud
MNC Cloud
PMN Cloud
Packed
RBC
After each category of cells was counted, the samples were
centrifuged (1200 x g, 10 mm) and their supernates were
carefully aspirated. We then added 3 mL of undiluted nitric
acid (Ultrex, J. T. Baker) to the tubes and heated the sample
at 135 #{176}C
in a block heater until only a residue remained.
This treatment with acid was repeated twice more. After the
samples were digested, we added exactly 1 mL of a 10 milL
solution of HNO3 to dissolve the mineral salts.
We diluted 1 mL of the erythrocyte suspension ninefold
with distilled de-iomzed water and analyzed this directly
without further digestion.
We determined the zinc content of the samples by atomic
absorption spectrophotometry
as adapted (16) for microsamples. A small Teflon cup, described by Manning (16),
was attached to the nebulizer of the atomic absorption
spectrophotometer,
which was set to record peak heights.
Four times we injected 100 .tL of sample into the cup with
an Eppendorf pipette. The first injection, used to rinse any
carryover from the previous sample, was not recorded. The
final three peak-height readings were averaged and the zinc
contents of the samples were calculated by comparison with
results for appropriate standards. To determine the precision and accuracy of the analyses, we concurrently assayed
in each run Bovine Liver standard (U.S. National Bureau of
Standards Standard Reference Material no. 1577), a plasma-pool sample, and samples to which a known quantity of
zinc had been added.
Data were evaluated
by analysis of variance, and the
significance of differences was examined by Scheff#{233}
contrasts (18).
Results and Discussion
Cell Separation
Fig. 1. Density gradient separation of blood cellular components
Platelets were resuspended in isotonic saline and centrifuged for 10 mm at 1200 x g. The supernate was decanted
and the platelets were resuspended in 11 mL of isotonic
saline. A 1-mL aliquot was then used for platelet counts.
The remaining sample was reserved for use in analysis for
zinc.
The MNC and PMN suspensions were diluted with repetitive equal volumes of isotonic saline, centrifuged (10 mm,
200 x g), then resuspended in 5 mL of isotonic saline. We
added 9 mL of distilled, de-ionized water and mixed the
sample by inversion for 15 a, then added 1 mL of a 90 g/L
solution of NaCl and mixed by inversion to restore isotonicity. We then centrifuged the sample at 200 x g and
discarded the supernate. This procedure was repeated one
to three times, as necessary, to remove any contaminating
erythrocytes. Once the pellet was free of erythrocytes we
resuspended it in 5 mL of isotonic saline and took an
aliquot for cell counts in a Coulter counter and for microscopic examination of the purity of the preparation.
We compared results by this procedure with those by
methods described by English and Anderson (14) as adapted
by Whitehouse et al. (10), who used a two-layered discontinuous gradient of Ficoll-Hypaque, and that of Jepsen and
Skottum (15), who used a two-layered discontinuous gradient of Percoll.
Erythrocytes were washed twice with isotonic saline and
66 CLINICAL CHEMISTRY, Vol. 31, No.
1, 1985
The method of separation and number of platelets remaining in the MNC fraction can influence the apparent
zinc content (Table 1). We found that published methods (14,
15) involving two-layered discontinuous gradient systems
inadequately separated platelets from MNC, and that the
three-layered discontinuous gradient that we describe here
separates the platelets from the MNC (Table 1). Platelets
account for about 70% of the zinc in the buffy coat (unpub-
Table 1. Effect of Separation Method and
Contamination with Platelets on Apparent Zinc
Content of Mononucleated Leukocytes (MNC)
Separation
method
Two gradient,
FicolI-
Sample
no.
ng
1 (3)C
2(3)
8.25
6.19
Zn/iD6
MNC
±
143d
±
0.94
Platelets
Corrected
remaining
ng Zn/b8
MNCb
4.63 ± 0.62
5.18 ± 1.08
after wash
194 ± 43
33
±
14
Hypaque
Two gradient,
Percoll
2 (3)
7.83 ± 0.74
76 ± 20
5.42 ± 0.34
Three gradient, 1 (2)
7.40 ± 0.81
37 ± 14 6.05 ± 0.18
Percoll with
2(3)
6.44 ± 1.69 3.4 ± 4.4 6.35 ± 1.69
NaCI wash
aTwo samples were split and run by different methods on separate
bApparent MNC Zn after correction for platelet contamination.
Plateletsin sample 1 contained 0.36 ng Zn/i 06 platelets; those in sample 2,
0.44 g Zn/i 06 platelets. Numberofreplicatesinparentheses.dMean± SD.
occasions.
lished observations), so a good separation from the MNC is
essential if the zinc content of MNC is to be accurately
estimated. Even if 90 to 95% of the platelets are removed, as
in the case of the two-gradient systems, there is sufficient
platelet zinc remaining in the preparation to significantly
increase the apparent zinc content of the MNC.
As others (10) have noted, clumping of cells during
separation and washing was a problem with some samples.
If clumping occurred, lower apparent cell counts were noted,
and the zinc content of the cells was overestimated. Thus it
was necessary to discard samples where clumping was
noted. Clumping occurred most frequently when we used
systems involving two gradients for separation, because the
platelets were not adequately separated from the cells. This
problem occurred much less often if the blood was diluted
with isotonic saline immediately after drawing and the cells
were separated within 30 mm by use of the three-gradient
Percoll system.
The choice of anticoagulant also appeared to influence cell
separation. There was less clumping and better separation
of MNC from platelets when EDTA was used as an anticoagulant than when either citrate or heparmn was used. Williams et al. (19) showed that platelets adhere to both MNC
and PMN when heparmnis used as an anticoagulant. This in
turn led to clumping of cells during centrifugation. However, there was less platelet binding to leukocytes in the
presence of EDTA. Additionally, Healy and Egan (20) have
recently demonstrated that the whole-blood platelet count
and mean platelet volume are significantly lower in citrated
blood than in EDTA-treated blood. Films (i.e., smears) of
citrated blood also showed considerable platelet clumping.
This suggests that the larger platelets aggregate in citrate.
These would most likely be associated with the cellular
fraction after centrifugation.
There was a good correlation between the DNA content
and cell counts of preparations obtained by use of the
present method. Evidently, clumping is not as great a
problem with this procedure as with earlier methods (10).
Microscopic
examination of the individual cell preparations indicated a good separation of the various cell species.
The platelet fraction was contaminated with fewer than 0.5
x 106 MNC per 1000 x i0 platelets. Fewer than 0.6% of the
platelets from the original sample remained in the MNC
fraction after separation. The MNC preparation also contained fewer than 1% PMN. The PMN preparation conmined 97-99% PMN and 1-3% MNC. No erythrocytes were
observed in this fraction after the lysing step. This procedure did not significantly change the apparent zinc content
of leukocytes. After separation and washing, approximately
80% of the platelets, between 60 and 70% of the MNC, and
70-80% of the PMN in the original whole-blood sample
could be accounted for.
The integrity of MNC and PMNs is maintained in the
samples separated over Percoll. In contrast, cells separated
by dextran sedimentation followed by gradient centrifugation over Ficoll-Hypaque (14) appeared to be deformed, and
yields were relatively poor. Others report (22) that neutrophils collected in this manner do not have normal sodiumtransport characteristics and may therefore have suffered
membrane damage. Dextran and Ficoll gradients subject
the cells to large osmotic stresses, whereas Percoll produces
gradients that are virtually iso-osmotic under physiological
conditions of ionic strength and osmolarity (12, 13).
1000 ng/mL, after which it began to flatten,as in convenFAAS (Figure 2). Most of our samples contained
between 100 and 500 ng of zinc per milliliter.
Peak heights were similar for irjected samples with equal
amounts of Zn but volumes of 100 or 200 L; such samples
gave disproportionately
smaller peak heights if their volume was <75 1iL.
The agreement was excellent between our results and
National Bureau of Standards certified values for a digested
Bovine Liver (SRM 1577) standard: we found 131.8 ± 5.5
(mean ± SD of 19 consecutive runs) g ofZn per gram dry
weight as compared with the certified Zn value of 130 ± 10
ug/g. Within-run and run-to-run CVs were 2.3% and 3.5%,
respectively, for 19 runs conducted over a three-month
period. Nearly all (98.5%, on the average) of the zinc added
to samples of cells before digestion could be accounted for
(Table 2).
tional
The micro-flame
atomic absorption spectrometric analysis
0.
LU
C-)
z
0
C,)
0.1
500
1000
ZINC CONCENTRATION
1500
(ng/mL)
Fig. 2. Standardcurves for zinc determination by injection FAAS (-)
and by regular aspiration FAAS (--------)
Table 2. Analytical Recovery
Zinc content, ng
Original
Platelets
176
176
Added
Av recovFound
ery, %
CV %
200
400
372(17)
578 (26)
200
282 (3)
98
1.0
200
363 (13)
97
3.5
98
101
4.6
4.6
MNC
87
Analysis for Zinc
PMN
The operating range for zinc analysis by this micro-flame
“sip” procedure was between 10 and 1000 ng of Zn per
milliliter of solution. The standard curve was linear up to
#{149}
Mean (and SD) of three replicates for each concentration of zinc added.
Zincwas added to the cells just before the digestion step.
170
CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985
67
for zinc offers several advantages over graphite-furnace
techniques.
Matrix
interferences
encountered
when the
graphite furnace is used are not a problem with the flame
technique (16, 17). Thus, standards need not be prepared in
an artificial
matrix, and standard curves are linear in the
operating range (10, 21). Most biological samples require
relatively large dilutions before readings will fall within the
operating range of the graphite furnace, 0-20 ng/mL. Most
of the samples that we examined contain 100-500 ng of Zn
per milliliter of sample. Thus, large dilutions are not
necessary with the micro-flame method, which has a much
broader operating range than the graphite furnace.
Table 3 shows the mean concentrations of zinc in cellular
components of blood from apparently healthy, free-living
adults. The population consisted of approximately equal
numbers of men and women, 25 to 49 years old. Most were
not supplementing their diet with Zn-containing preparations. The zinc in platelets, MNC, PMN, or erythrocytes did
not appear to be greater in the three individuals who were
taking mineral supplements than in those who were not.
Our values for zinc-7.4 ± 2.3 and 5.1 ± 1.1 ng/106 cells
for MNC and PMN, respectively-are
lower than those
reported previously (Table 4). In most earlier reports the
contribution of platelet zinc, particularly to the MNC fraction, was not considered. As noted earlier (Table 1), significant numbers of platelets remain in the MNC fraction, even
after washing, when two-gradient separation systems are
used. The higher percentage of platelets removed when a
three-gradient system of Percoll is used, as we described,
results in the lower apparent zinc content of the MNC. Our
values of 0.48 ng/106 for platelets and 1.13 ng/106 or 38.7
j.&g/g of hemoglobin for erythrocytes agree well with those
found by others (10, 23, 28).
It remains to be shown whether determining the Zn
content of any of these categories of cells will provide a more
reliable index to whole-body zinc nutrition. This method,
Table 3. Zinc Content of Various Cellular
Components of Blood from Healthy, Free-living
Adults
No.
determinations
Zn, ng/10 celis
0.48 ± 0.18
7.4 ± 2.3
Platelets
MNC
PMN
Erythrocytes
5.1
1.13
±
±
43
45
39
81
81
1.1
0.10
38.7 ±
1284 ± 125c
29b
Mean
± SD.
g/g
81
of hemoglobin. cmg/L of packed cells.
Table 4. Comparison of Reported “Normal”
Values for Zinc Content of Human
Leukocyte Populations
Zn, ng/10 cells
Leukocytes
32.0 ± 13.08
10.3 ±
11.2±
13.7±
8.0
±
2.5
1.9
4.0
MNC
no.
24
25
18.8
11.5
±
3.0
8.6
±
1.5
10.4
±
2.3
1.2
26
±
10.9
±
2.6
2.3
4.9
±
1.3
5.1
±
1.1
27
This study
10
2.2
7.4
Mean
PMN
±
± SD.
68 CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985
results of which are described here only for normal subjects,
should now be tried in patients with various forms of illness,
including those whose zinc nutriture
is believed to be
abnormal.
The competent
technical
assistance
of Patrick
Theisen,
Kathy
Huot, and Eugene Korynta is gratefully acknowledged.
References
1. Underwood EJ. Trace Elements in Human and Animal Nutrition, 4th ed., Academic Press, New York, NY, 1977, pp 196-242.
2. Prasad AS (Ed.). Trace Elements in Human Health and Disease,
VI, Zinc and Copper, Academic Press, New York, NY, 1976.
3. Sandstead HH. Zinc nutrition in the United States. Am J Clin
Nutr 26, 1251-1260 (1973).
4. Casey CE, Hambridge KM. Epidemiological aspect of human
zinc deficiency. In Zinc in the Environment, Part II: Health Effects,
JO Nriagu, Ed., John Wiley & Sons, New York, NY, 1980, pp 1-27.
5. Solomons NW. On the assessmentof zinc and copper nutriture in
man. Am J Clin Nutr 32, 856-874 (1979).
6. Beisel WR, Pekarek ES. Acute stress and trace element metabolism. In Neurobiology of the Trace Metals Zinc and Copper, CC
Pfeiffer, Ed., Academic Press, New York, NY, 1972, pp 53-82.
7. Prasad AS, Rabbani P, Abbasii A, et al. Experimental
zinc
deficiency in humans. Ann Intern Med 89, 483-490 (1978).
8. Jones RB, Keeling PWN, Hilton PJ, Thompson RP. The relationship between leucocyte and muscle zinc in health and disease. Clin
Sci 60, 237-239 (1981).
9. Meadows NJ, Ruse W, Smith MF, et al. Zinc and small babies.
Lancet ii, 1135-1137 (1981).
10. Whitehouse RC, Prasad AS, Rabbani P1, Cossack ZT. Zinc in
plasma, neutrophils, lymphocytes and erythrocytes as determined
by flameless atomic absorption spectrophotometry.
Clin Chem 28,
475-480 (1982).
11. Hinks LI, Colmsee M, Delves lIT. Determination
of zinc and
copper in isolated leucocytes. Analyst 107, 815-823 (1982).
12. Pertoft H, Laurent TO. Isopyrenic separation of cells and cell
organelles by centrifugation
in modified colloidal silica. In Methods
in Cell Separation, N Catsimpoolas, Ed., Plenum Press, New York,
NY, 1977, pp 25-65.
13. Pertoft H, Laurent TO, Seljelid R, et al. The use of density
gradients of Percoll for the separation of biologicalparticles. In
Separation of Cells and Subcellular Elements, H Peeters, Ed.,
PergamonPress, New York, NY, 1980, pp 67-72.
14. English D, Anderson BR. Single step separation of red blood
cells, granulocytes
and mononuclearleucocytes on discontinuous
gradients of Ficoll-Hypaque.
J Immunol Methods 5, 249-252
(1974).
15. Jepsen LV, Skottum TA. A rapid one-step method for the
isolation of human granulocytes from whole blood. Scand J Clin
Lab Invest 42, 235-238 (1982).
16. Manning DC. Aspirating small volume samples in flame atomic absorption spectroscopy. At Abs News 14, 99-102 (1975). (A
Perkin-Elmer
publication.)
17. Slavin M. Atomic Absorption Spectroscopy, 2nd ed., John Wiley,
New York, NY, 1978.
18. Scheff#{233}
H. The Analysis of Variance, John Wiley, New York,
NY, 1959, pp 90-137.
19. Williams WR, Cavolina F, Williams WJ. Platelet binding in
mononuclear
cell preparation
from peripheral
blood. Br J Exp
Pathol 64, 451-455 (1983).
20. Healy DT, Egan EL. Centrifugal
and anticoagulant induced
variations in platelet rich plasma and their influence on platelet
aggregation. Scand J Haematol 32, 452-456 (1984).
21. Viera NE, Hansen JW. Zinc determination
in 10-giL serum or
urine samples by fiameless atomic absorption spectrometry.
Olin
#{231}hem
27, 73-77 (1981).
22. Poston L, Jones RB, Hilton PJ. Sodium transport in polymorphonuclear leucocytes: Effect of isolation by the FicolVFriosil method. Clin Sci 62, 563-564 (1982).
_________
23. Foley B, Johnson SA, Hackley B, et al. Zinc content of human
platelets. Proc Soc Exp Biol Med 128,265-269(1968).
24 Vallee BL, Gibson JG. The zinc content of normal human whole
blood leucocytes and erythrocytes. J Biol Chem 176, 445-447
(1948).
25. Fredricks RE, Tanaka KR, Valentine WN, Mattson M. Zinc in
human blood cells: Normal values and abnormalities associated
with liver disease. J Olin Invest 39, 1651-1656 (1960).
26; Nishi Y, Hatano S, Horino N, et al. Zinc concentration in
leucocytes: Mononoclear cells, granulocytes, T-lymphocytes, T-lymphocytes and monocytes. Hiroshima J Med Sci 30, 65-69 (1981).
27. Purcell SK, Jacobs MA, Hambridge KM. Neutrophil and mononuclear leucocyte zinc. Fed Proc Fed Am Soc Exp Biol 43, 687
(1984). Abstract.
28. Kiem J, Borberg H, Iyengar GV, et al. Elemental composition
of platelets. Part II. Water content of normal human platelets and
measurements of their concentrations of Cu, Fe, K, and Zn by
neutron activation analysis. Olin Chem 25, 705-710 (1979).
CLINICALCHEMISTRY, Vol. 31, No. 1, 1985 69