CLIN. CHEM. 30/2, 271-274 (1984)
Methods for Extracting Phospholipids from Human Amniotic Fluid Compared
C. G. Duck-Chong, G. J. Baker, S. R. Murdoch, and R. M. Price
Phospholipids were extracted from human amniotic fluid by
various procedures, including the two most commonly applied to amniotic fluid for evaluation of fetal lung maturity. We
find
that
the
yield
of phospholipid
is greatly
procedure
dependent.This should be taken into account when one is
consideringthe various reported methods of evaluating fetal
lung maturity, because in some of them phospholipid data
are expressed in terms of absolute concentration in the
amniotic fluid. There were also significant differences in
phospholipid composition in extracts prepared by the various
procedures, but in general these were not large enough to
influence evaluation of fetal lung maturity by methods in
which phospholipid data are expressed in relative terms, as
ratios or percentages-e.g.,
in the Iecithin/sphingomyelin
ratio and “lung profile” procedures. In the extraction method
originally recommended for determination of the lecithin!
sphingomyelin ratio, both the yield and composition of phospholipid
depend on the extent of mixing.
Additional Keyphrases: fetal status
drome
analytical error
respiratory distress syn-
As the fetal lung matures, lecithin-rich pulmonary surfacaccumulates in the amniotic fluid. Extraction and
determination of phospholipids in amniotic fluid therefore
forms the basis of many of the methods proposed for clinical
evaluation of fetal lung maturity (1). Since 1971, when the
first such method was introduced (2), many potential
sources of within- and between-laboratory variation in the
determination
of amniotic fluid phospholipids have been
recognized
(1). The use of different methods for the initial
extraction of phospholipids from the sample has been suggested as a possible source of variation (1), but no published
data are available for evaluation of this variable.
In the procedures of Foich et al. (3) and Bligh and Dyer
(4), both commonly used in research laboratories, phospholipids are initially extracted with appropriate large volumes
of chloroform and methanol to give a single phase, which
then is separated into two phases by increasing the water
content. Lipids are extracted into the chloroform (lower)
phase and non-lipid impurities
remain in the aqueous
tant
(upper)
phase. The Folch procedure
also includes
filtration
of the single-phase extract, to remove precipitated proteins.
These procedures were carefully designed to minimize problems inherent in the extraction of lipids: incomplete extraction, loss of specific phospholipid classes by adsorption to
protein, and loss of part of the more water-soluble lipids into
the aqueous phase. Alternatively, the last problem may be
overcome by chromatographing the initial single-phase exDepartment
of Histology and Embryology,University of Sydney,
Sydney, N.S.W., Australia 2006.
Received July 6, 1983; accepted October 18, 1983.
tract on a column of Sephadex (5, 6).
In contrast, in clinical laboratories phospholipids are
usuallyextracted
by mixing a small volume of amniotic
fluid with an equal volume of methanol, then vortex-mixing
with one or two volumes of chloroform, which yields in one
stepa biphasicmixture with phospholipid in the chloroform
(lower) layer. The latter approach has been widely adopted,
because it is rapid and solvent volumes are conveniently
small.
In this study, we compared lipid extracts prepared from
amniotic fluid by various clinical and research methods with
respect to the yield of phospholipid, the relative proportions
of the different phospholipid classes in the extract, and (for
one method) the effect of varying the extent of mixing
during extraction.
Materials and Methods
Samples of amniotic fluid were obtained by transabdominal amniocentesis as part of routine management of complicated pregnancies. After centrifugation, a portion of the
supernate was used without delay for determination of the
lecithinlsphingomyelin
(uS) ratio (7), and the rest was
stored at -20 OCI Samples with L’S ratios of 2 to 5.5 were
selected for use in this study. Details of many of the reagents
and materials used are given elsewhere (8, 9).
Extraction Methods
fluid was extracted at
20 volumes of a 2/1 by volume mixture of
chloroform and methanol. We used the procedure described
by Nelson (5) for plasma. After it had stood for 30 mm, the
extract was filtered through a firm plug of glass wool, which
was washed with a little more chloroform/methanol.
Nonlipid contaminants were removed by either of the following
methods: (a) Folch-wash. A portion of the extract was
shaken with 0.2 vol of isotonic saline and the lower phase
was collected when it became clear (3). (b) Chromatography
on Sephadex. A portion of the extract containing about 40
g of P was dried2 and the residue extracted
with 5 mL of
Folch lower phase (FLP). The residue was then dispersed in
about 1 mL of Folch upper phase (FUP) to disperse any
entrapped lipid (6), dried, and the residue extracted with a
further 5 mL of FLP. The combined FLP extracts were
dried, redissolved in a small volume of FLP, and applied to a
0.5 x 7 cm column of Sephadex G25, coarse (Pharmacia,
Method
of Foich et al. (3). Amniotic
0 #{176}C
with
1 Nonstandard
abbreviations: L/S, lecithinfsphingomyelin;
FLP,
Folch lower phase, that is, chloroform/methanollwater
(86/14/1, by
vol); FUP, Folch upper phase, that is, chloroform/methanol/water
(3/48/47,by vol); and PC, FE, P1, LL, PG, SM, see Table 1.
2”Drying” (i.e., removal of solvents) was routinely done by rotary
evaporation at 25#{176}C.
For samples with a high water content, small
volumes of chloroform/methanol 1211by vol) were added repeatedly
to facilitate evaporation.
CLINICAL CHEMISTRY. Vol. 30. No. 2, 1984
271
Sweden), which had been washed with FUP and
equilibrated with FLP just before use (6). Phospholipids
were eluted with 10 mL of FLP, then non-lipid P was eluted
with 10 mL of FUP.
Method of Bligh and Dyer (4).Methanol,then chloroform,
was added to amnioticfluidto givea mixture consisting
of
chloroformlmethanollwater
(1/2/0.8
by vol).Aftervigorous
Uppsala,
mixing
and standing
for 2 mm, sufficient chloroform
and
then distilled water was added (with mixing aftereach
addition)
to bringthe proportionsofchloroform,methanol,
and water to 2/2/1.8.
The phases were allowed to separate
and most of the lower layer was collected. The rest was
collected after centrifuging the emulsion remaining at the
interface.
Method of Gluck et al. (2). (a) Replicate 2-mL portions of
amniotic fluid in 15-mL centrifuge tubes were vortex-mixed
for10 s with 2 mL of methanol. Then 4 mL of chloroform
was added to each tube and the mixture was vortexed again
(for a further 20s, unless otherwise indicated). After centrifugation, the clear lower phase was collected with a Pasteur
pipette. However, emulsion at the interface prevented complete collection. Therefore, blanks containing NaC1 (9 g/L)
instead of amniotic fluid were processed in the same way to
provide an estimate of the total volume of the lower phase in
the absence of any emulsion.(b) Same as for a, but only 2
mL of chloroform was used per tube.
Extracts obtained by all methods were finally dried and
redissolved in FLP. Total phospholipid P was determined in
triplicate (9).
Statistical
Analysis
Data were analyzed by analysis of variance, with comparison of means based on a t value adjusted (accordingto
Scheff#{233})
to allow for multiple comparisons, or by Student’s ttest (10). For each experiment, the phospholipid composition
was determined in batches consisting of one extract prepared by each of the methods so that data could be “paired”
accordingto assaybatch,if required.
Results and Discussion
Purifying the lipid extract by Sephadex chromatography
gave about a 15 to 20% lower yield of total phospholipid
than Folch-washing (Table 1).Incomplete extractionof
phospholipids
from the insolubleresidueof salts and proteinsthatwas leftafter evaporation of the original extract
before chromatography probably accounts for this result.
The difference cannot be attributed to less effective removal
of impurities by the Folch-wash method, since amniotic fluid
extracts contained only about 3% non-lipidP-that is,P
eluted from the column by FUP.
There were significant differences in the phospholipid
composition of the extracts obtained by these two methods
(Table 1). Folch-washing reduced the relative proportion of
lysolecithin (LL) by about 85% and that of phosphatidylethanolamine (PE) by 20-30%. The latter result is significant
for pooled data from batches 1 and 2 (p < 0.05, paired t- test).
Moreover, a tendency for the proportion of phosphatidylglycerol (PG) to be lower and phosphatidylinositol
(P1) to be
higher in the Folch-washed extracts resulted in a signifiDetermination
of Phospholipid
Composition
cantly lower PGIPI ratio in such extracts (p < 0.005, paired
Phospholipid classes were separated by two-dimensional
t-test, pooled data).
All the methods in Table 2 are related in the sense that, in
thin-layer chromatography on plastic-backed plates as described previously (8). Plates were prewashed by chromatoeach case, the lipid extract is “washed” by equilibration
graphing overnight in methanol/water (9/1 by vol). After
against an aqueous phase. The Gluck methods (a and b)
yielded about 40% less total phospholipid than did the other
application of the sample and between solvents the plate
was dried for 5 mm in a fume hood, then for 5 mm at 60 #{176}C.two methods (Table 2). Part, but not all, of this loss could be
Lipid extract corresponding to 4-5 g of P was dried in a
accounted for by entrapment of some of the chloroform layer
sample tube and transferred to the chromatographic plate
in the emulsion at the interface between the phases (see
with FLP. The total P content of any residue in the tube was
Footnote b, Table 2). A tendency for the Gluck methods to
give a higher proportion of sphingomyelin (SM) (and theredetermined (usually <5% of the total sample P). After
chromatography,
the plates were dipped in a modified
fore a lower PC/SM ratio) was noted and subsequently
molybdenum blue reagent (8).
confirmed (Table 2).
The value for PG given by the Bligh and Dyer method
As soon as the plates appeared dry, the blue phospholipid
spots were cut out with scissors, then held with fine forceps
tended to be lower than that given by the other methods
(Table 2), but the difference was not statistically significant.
over a test tube while the silica gel was washed from the
backing with 2-3 mL of chloroform/methanol
(2/1 by vol)
However, because information
on the amount of it in
directed in a gentle stream from a Pasteur pipette. (The
amniotic fluid can be useful in assessing fetal lung status
(11), we investigated
this point further. An additional
silica gel should slide off the plate in one piece and drop to
the bottom of the tube without touching the sides.) Blank
sample of amniotic fluid was extracted by the Bligh and
areas of the plate were treated in the same way.
Dyer and the Folch-wash methods. In this experiment
the
Magnesium nitrate was added and phospholipid P was
mean percentage of PG in the extract (mean ± range, n = 3)
determined as described previously (9), with the following
was 2.6 ± 0.2 and 2.6 ± 0.3 for the Bligh and Dyer and the
modifications: (a) After the addition of HCI and hydrolysis,
Foich-wash methods, respectively. It therefore seems unthe tubes were allowed to stand until the silica gel had
likely thatthere is any real difference in the PG content of
settled;
the colorreactionwas then carriedout on 0.5 mL of
the extracts obtained by the different extraction methods.
the supernate with use of half the usual volumes of Triton
More LL was recovered by the Gluck methods (a and b)
X-l0O and malachite green reagent. (b) For phosphatidylthan by the Bligh and Dyer and Folch-wash methods (Table
2), but the valueobtainedwas still
considerably
lower than
choline (PC), six times the usual volumes of Mg(N03)2 and
HC1 were used and a 0.3-mL sample, diluted to 0.5 mL with
that predicted for the Sephadex method by the data in Table
HC1, was used for the color reaction.
1. Using the latter method, we found LL to be present in
The whole procedure was carried out in duplicate or
amniotic fluid in amounts comparable to SM, PG, and P1
triplicate for each extract. The P content of each phospholip(Table 1). In many previous investigations, the amount of
LL in amniotic fluid has probably been grossly underestiid spot was expressed as a percentage of the total P
mated because extraction methods involving partitioning
recovered in all spots. The total P in the phospholipid spots
against an aqueous phase have been used.
togetherwith the residue in the sample tube usually acFor the Gluck method a, the method originally recomcounted for about 80% of the total P in the sample (mean
mended (2) and still widely used in clinical laboratories in
79%, SEM 2%; n
23).
=
272 CLINICAL CHEMISTRY,Vol. 30, No. 2, 1984
Table 1. Results Compared for Two Methods for Removing NonhipidContaminants from Amniotic Fluid
Lipid Extractsa
Batch 2
Batch I
Method
I.Total PL,b mg/L
PA,%
PG,%
PE,%
PS,%
PI,%
PC,%
SM,%
LL,%
PC/SM
Sephadex
134 ±1
3.9 ± 0.4
4.3 ± 0.1
2.8± 0.3
1.6 ± 0.3
8.4 ± 0.2
72.3 ± 1.1
4.2± 0.2
2.8± 0.3
17.5 ± 0.9
Foich-wash
158
+4**C
Sephadex
92
±1
1.4 ± 0.8
4.1 ± 0.9
2.8 ± 0.1
1.2 ± 0.2
1.9 ± 0.1
3.0± 0.5
3.5 ± 0.2
2.9± 0.8
9.4± 1.0
74.8 ± 1.2
4.1 ± 0.2
0.4 ± 0.1#{176}’
19.5 ± 0.4
0.3 ± 0
6.1 ± 0.6
75.3 ± 1.6
4.1 ± 0.2
5.0
± 0.1
Foich-wash
112
±1”
± 0.4
± 0.3
± 0.5
± 0.5
7.1 ± 0.5
76.0 ± 1.0
4.4± 0.4
0.7 ± 0.5#{176}’
18.0 ± 2.0
0.5 ± 0’
2.2
3.3
4.0
2.3
18.5 ± 0.9
PG/PI
0.5 ± 0
0.7 ± 0.1
‘Two differentbatches of pooledamniotic fluid wereextracted;single samples(batch 1)or duplicatesamples(batch2) were processedby each of the alternative
methods. Fortotalphospholipid, eachvalue representsthe mean ± SEM of triplicate determinationsfor each extract. For phospholipidcomposition, each value
represents
themean ± range of two determinations(batch 1) or ± SEM of five determinations (batch 2).
bpL, phospholipid (phospholipid P x 25); PA. phosphatidic acid; PC, phosphatidylglycerol; PE, phosphatidylethanolamine; PS, phosphatidylserine; P1,
phosphatidylinositol;
PC, phosphatidylcholine;
SM, sphingomyelin; LL, lysolecithin.
CA valuemarked with asterisks is significantlydifferent fromthe correspondingvalue given by the Sephadexmethodfor the same batch: p < 0.05; p < 0.025
(paired t-test).
Table 2. Comparison of the Phosphohipid Composition of Lipid Extracts Obtained from Amniotic Fluid
by Four Different Procedures a
Folch-wash
I.Total PL, mg/L
PG,%
PE,%
PS,%
Pl,%
PC,%
SM,%
LL,%
II. PC,%
73 ±2
3.5 ± 0.3
4.2 ± 0.2
4.1 ± 0.5
8.3 ± 0.4
73.6 ± 0.7
5.9 ± 0.3
0.4 ± 0.2
73.7± 0.5
5.2± 0.5
BIIgh-Dy.r
77 ±1
2.6± 0.3
4.1 ± 0.5
3.9± 0.4
8.1 ± 0.7
75.5 ± 1.0
5.5± 0.3
0.4± 0.2
75.5± 0.7
Gluck a
43
±1k’
3.2± 0.2
4.4 ± 0.3
4.0 ± 0.4
7.5 ± 0.5
73.3± 0.8
6.5± 0.5
1.1 ± 0.2c
74.0 ± 0.7
5.8± 0.5
Gluck b
45 ±
3.5 ±
4.6 ±
2.9 ±
2
0.2
0.4
0.3
0.3
7.7 ±
± 1.1
± 0.4
± 0.2c
74.1 ± 0.9
6.1 ± 0.4c
12 ±1C
74.0
6.7
0.7
5.1 ± 1.7
15 ±1
15
±1
13 ±1C
‘For I, duplicateextractionswerecarried out inparallelby the methodsindicated ona pooledsampleof amnioticfluid, and the phospholipid composition of each
extractwas determinedby two-dimensional thin-layer chromatography. For II, a second pooled sample of amniotic fluid was extracted by each method and
phospholipid composition was determined only by one-dimensionalthin-layer chromatography.These data were pooled with the data in I for statistical analysis.
Eachvaluerepresentsthe mean ± SEMof six (I) or nine (II) determinations.For key to abbreviations,seeTable 1.
b Beforecorrecting
forincompleterecovery of the lower layer (see Materials and Methods). Aftercorrection,the yields of total phospholipidwere 58 and 62 mg/L
for methods a and b, respectively.
CThe overall mean value for the two Cluck methods differs significantly from the overall mean value given by the other two methods (p < 0.05, analysis of
SM,%
PC/SM
variance).
determiningthe L/S ratio,resultsobtainedby vortex-mixingfor either 15 or 30s were generally similar, although the
significantly
lower yield of PG given by the shorter mixing
time is noteworthy (Table 3). Mixing by sharply inverting
the tube instead of vortexing resulted in a lower yield of
total phospholipid (only about 60% of the value predicted for
the Folch-wash procedure by the data in Table 2). Inverting
gently instead of sharply, a procedure recommended in
another investigation (12), reduced the yield from 60% to
30% (data not shown).
When the sample was mixed by inversion, phospholipid
percentage composition and PC/SM ratio were both significantly altered (Table 3). Moreover,in experiment II(Table
3), an overall correlation was found between the PC/SM
ratio of the individual
extracts and the yield of total
phospholipid obtained (p < 0.01, n = 12); the latter value
presumably reflects the degree of mixing achieved during
extraction of a particular sample, irrespective of the mixing
procedure used. These findings indicate that the intensity of
mixing during extraction of lipid from amniotic fluid represents another potential source of between-operator and
between-laboratory
variation
in the assessment of fetal lung
maturity by the L/S ratio method. The significantly lower
PG/SM ratio given by the less-vigorous mixing procedures
(Table 3, experiment I) suggests that the extent of mixing
may also be critical
in procedures for assessing fetal lung
maturitywhich requiremeasurement ofPG. When biphasic
extraction methods are used, samples should not be mixed
by inversion.Rather, they should be vortexed thoroughly
under well-defined and reproducible conditions, as previously recommended (13).
Of particular interest is the finding that the commonly
used clinical methods (Gluck methods a and b) yielded
considerably less phospholipid than the other methods (Table 3). The yield obtained by the former methods is likely to
be quite variable because it depends on both the extent of
mixing (Table 3) and the volume of the lower layer trapped
in the emulsion remaining at the interface after centrifugation. The choice of extraction method is therefore an important factor to be borne in mind when reviewing the literature relating to all methods for evaluating fetal lung maturity in which phospholipid data are expressed in terms of
concentration
in the amniotic fluid. These include methods
based on the determination of total phospholipid, lecithin,
disaturated
lecithin, palmitic acid, “10,000 g pellet” phospholipid, and “lamellar body” phospholipid (1).
CLINICAL
CHEMISTRY,
Vol. 30, No. 2, 1984
273
Table 3. The Dependence of Yield and Percentage
Composition of Phosphohipid on the Extent of
Mixing during the Extraction of Amniotic Fluid by
the Gluck Methoda
Procedure
Vortex 30 $
I. TotalPL,#{176}
mg/L
PG, %
PE. %
PS, %
P1,%
PC, %
SM, %
PC/SM
PG/SM
II. Total PL,
mg/L
“PC/SM”
74
4.6
4.9
2.3
6.7
75.6
±
±
±
±
±
±
±
1
0.1
Vortex 15 a
73
4.0
0.5
4.2
0.2
2.5
0.3
6.7
0.5
6.7
0.1
11.3 ±0.2
0.69 ± 0.02”
76.0
6.7
55
3.3
52
3.1
±
±
3
0.2
± 2
± 0.1***
± 0.1
± 0.2
± 0.3
± 0.9
± 0.2
11.4 ±0.5
0.60 ± 0.01
± 2
± 0.1
Invert
53
4.8
4.2
3.0
8.4
71.9
77
±
±
±
±
±
±
±
9.4 ±
0.63 ±
36
2.7
±
±
1”#{176}
0.1
0.1
0.1’
0.6*
1.0#{176}”
0.6*
0.03
3***
‘Replicate
2-mL samples of pooled amniotic fluid were extracted by the
Cluck method a. Afterchloroformwas addedthe mixture was either vortexmixed for 15 or 30 S or mixed by inverting the stoppered tube sharply eight to
10 times.InexperimentI,six2-mL samples were processed by each of the
vortexing procedures and nine by inversion. For each mixing procedure, the
chloroform layers were pooled to provide a single extract for the determination
of total phospholipid and phospholipid composition. Each value representsthe
mean ± SEM of three determinations. In experiment II, four samples of
amnioticfluidwere extractedby each procedureand the chloroformlayers
were collectedindividually.Total phospholipid was determinedin duplicate.
then phospholipidswere separatedbyone-dimensional
thin-layerchromatography of duplicate samples, each containing 0.7 g of P. The PC/SM ratios
weredetermined densitometricallyafter staining with cupric acetate-phosphoric acid reagent (7); in this experiment the PC/SM ratio is equivalent to the (iS
ratiousedfor clinicalassessment,not a molarratioas in experimentI. Each
valuerepresentsthemean ± SEM of the four results obtained for each mixing
procedure.
0For key to abbreviations, see Table 1. Total PL values have been
corrected for incomplete recovery of the lower layer.
A value marked with asterisks is significantly different from the overall
meanvaluegiven by the other two procedures:*p < 0.01, p < 0.025, “p <
0.005.
the variation from this source is not generally large
enough to cause problems in the evaluation of fetal lung
maturity.
In such cases, the convenient single-step methods
now in use appear to be quite satisfactory
if biphasic
mixtures
of chloroform, methanol, and amniotic fluid are
vortexedadequatelyduring extraction.
practice
This work was supported by grants from the National Health and
Medical Research Council
of Australia, Royal North Shore Hospital
and the Utah Foundation.
We are also grateful to Dr. W. J.
Hensley, Royal Prince Alfred Hospital,
Carnperdown, N.S.W., for
providing amniotic fluid.
References
1. Brown LM, Duck-Chong CG. Methods for evaluating fetal lung
maturity. CRC Crit Rev Clin Lab Sci 16, 85-159 (1982).
2. Gluck L, Kulovich MV, Borer RC, et al. Diagnosis of the
respiratory distress syndrome by amniocentesis. Am J Obstet
Gynecol 109, 440-445 (1971).
3. FolchJ,Lees M, Sloane-Stanley GH. A simple method for the
isolation and purification of total lipids from animal tissues. J Biol
Chem 226, 497-509 (1957).
4. Bligh EG, Dyer WJ. A rapid method of total lipid extraction and
purification. Can J Bi±ichem Physiol 37, 911-917 (1959).
5.Nelson GJ. In Analysis of Lipids and Lipoproteins, EG Perkrns,
Ed., American Oil Chemists’ Soc., Champaign, IL, 1975, pp 1-21.
6. Radin NS. In Methods in Enzymology, 14, SP Colowick, NO
Kaplan, Eds., Academic Press,New York, NY, 1969, pp 245-254.
7. Brown LM, Duck-Chong CG, Hensley WJ. Improved procedure
for lecithin/sphingomyelin
ratio in amniotic fluid reduces false
predictions of fetal lung immaturity. Clin Chem 28,344-348(1982).
8. Duck-Chong CG, Baker GJ. Rapid qualitative analysis of polar
and non-polar lipids in a single sample using “three-way” thin layer
chromatography. Lipids 18, 387-389 (1983).
9. Duck-Chong CG. A rapid sensitive method for determining
phospholipid phosphorus involving digestion with magnesium nitrate. Lipids 14, 492-497 (1979).
10. SnedecorGW, Cochran WG. Statistical Methods, 6th ed., Iowa
State University
In other methods, such as the L/S ratio (2) and the lung
profile (11), phospholipid data are expressed in relative
terms as ratios or percentages. While the present study has
revealed that the results of such procedures are influenced
to some extent by the choice of extraction method, in
274
CLINICAL CHEMISTRY, Vol. 30, No. 2. 1984
Press, Ames, IA, 1969.
11. Kulovich MV, Hallman MB, Gluck L. The lung profile.
I.
Normal pregnancy. Am J Obstet Gynecol 135, 57-63 (1979).
12. Johnson LW. Determiningfetal lung maturity: A sensitive
surfactant method. Am J Obstet Gynecol 129, 190-196 (1977).
13.Gluck L, Kulovich MV, Borer RC. Estimates of fetal lung
maturity. Clin Perinatol
1, 125-139 (1974).