Microparticle-enhanced nephelometric immunoassay of lysozyme in milk and other human body fluids
Author for correspondence. Fax 33 383 44 60 22; e-mail
This immunoassay has a detection limit of 8 µg/L of reaction mixture and can be performed using diluted milk (1:6000, in reaction mixture), excluding sample pretreatment. Human milk lysozyme can be quantified over the concentration range 0.091.50 g/L, with within- and between-run coefficients of variation <5%.
Changes in the lysozyme concentration of human milk during lactation were determined in 636 samples. Lysozyme concentrations (mean ± SE) decreased from colostrum (0.36 ± 0.02 g/L) to transitional milk (0.30 ± 0.01 g/L) and mature milk during days 1542 (0.30 ± 0.01 g/L), then increased in the mature milk during days 4356 (0.35 ± 0.01 g/L) and especially during days 5784 (0.83 ± 0.05 g/L). The proportion of lysozyme contributing to total protein was found to rise during lactation and was as follows: colostrum (1.7%), transitional milk (2.3%), and mature milk from days 1528 (2.7%), days 2942 (3.1%), days 4356 (3.8%), and days 5784 (7.3%).
The assay developed for milk was also suitable for the determination of lysozyme in other human body fluids.
However, LZ is a minor whey component of human milk compared with secretory IgA, lactoferrin, and -lactalbumin (5). LZ is a glycosyl hydrolase, cleaving the ß-1,4 bonds between N-acetylglucosamine and N-acetylmuramic acid (6). By hydrolyzing the peptidoglycans of procaryote cell walls, human milk LZ has a bacteriolytic function and plays a role, together with secretory IgA and lactoferrin, in the passive protection of breast-fed newborns (7)(8).
Human milk LZ and -lactalbumin are also evolutionarily related, with conservation of the position of four disulfide bonds and of 40% of amino acid residues (9). Immunoassays of milk LZ should use antibodies that do not cross-react with -lactalbumin.
Several methods of quantitation of LZ in milk have been reported, such as fast protein liquid chromatography (10), polyacrylamide gel electrophoresis (11), enzyme activity assessment (12)(13), and immunoassays (14)(15)(16). Enzyme activity, inducing the lysis of Micrococcus lysodeikticus, is commonly measured in human body fluids by turbidimetric (17), nephelometric (18), and lysoplate (19) techniques. Immunochemical quantitation of LZ uses immunoelectrophoresis (20), radioimmunoassay (21), enzyme-linked immunosorbent assay (22), and classical immunonephelometry (23)(24).
As previously pointed out (14)(17)(18)(22)(23)(24), most of these methods are lacking in sensitivity and require pretreatment of the samples and, sometimes, a long incubation period.
Microparticle-enhanced nephelometric immunoassays have been previously described as sensitive and accurate techniques for the determination of various human serum proteins (25) and of the main components of bovine milk (26). Such immunoassays are based on the nephelometric quantification of the inhibition of microparticleantigen conjugate immunoagglutination by the antigen to be assayed (27).
Microparticle-enhanced nephelometric immunoassays have been applied more recently to the measurement of -lactalbumin (28), lactoferrin (29), and ß-casein (30) in human milk. In such applications, they present the advantages of requiring no pretreatment other than high dilution and of using a single technique for assaying the main proteins of the complex medium that is human milk.
Here, we report the development of a microparticle-enhanced nephelometric immunoassay of LZ in human milk and its application to the investigation of the quantitative changes in milk LZ during the first 12 weeks of lactation.
The method assessed in milk was then applied to the determination of LZ in other human body fluids (serum, saliva, and urine). Finally, the analytical performances of the microparticle-enhanced nephelometric immunoassay and of the most frequently used other LZ assays were compared.
The specific reaction against human LZ was ascertained using crossed immunoelectrophoresis, which revealed that the LZ precipitation arc appears only against concentrated urine from patients with monocytic leukemia and that no precipitation is observed against nondiseased human plasma or concentrated nondiseased human urine.
Polyfunctional hydrophilic microspheres (MSs) of 300 ± 12 nm diameter, synthesized as previously reported (31), were supplied by Sanofi-Diagnostics-Pasteur. They were covalently coated with human LZ through the formation of imine bonds (27) between the amino groups of LZ (0.3 g/L) and the aldehyde groups on MSs (10 g/L).
After unreacted aldehyde groups were blocked, the MSLZ conjugate was recovered by centrifugation and stored at -20 °C, at 3.3 g/L, in a sucrose solution (50 g/L) supplemented with polyethylene glycol 6000 (50 g/L) and NaN3 (2 g/L).
Human milk samples used for evaluation of immunoassay development and for investigation of the changes in milk LZ concentration during lactation were collected from 74 mothers, who had volunteered, at the maternity hospital of Nancy (France) or at home.
The 636 samples were colostrum from days 15 postpartum (168 samples), transitional milk obtained days 614 (182 samples), and 286 mature milk samples obtained days 1528 (167 samples), days 2942 (63 samples), days 4356 (28 samples), and days 5784 postpartum (28 samples).
All milk samples were frozen immediately after collection and stored at -20 °C until use. They were thawed at 40 °C in a water bath and vigorously homogenized immediately before analysis. Total protein concentration in these human milk samples was determined by the Bradford method (32).
Serum, saliva, and urine samples were obtained from healthy subjects (1860 years old) and stored at -20 °C until assay. Blood (n = 30), randomly chosen from samples assayed for other purposes, was allowed to coagulate. Whole unstimulated saliva (n = 30) was collected by the draining method (33), then centrifuged for 3 min at 1000g. Urine samples (n = 50) were prepared by centrifugation for 15 min at 3000g.
Assay of milk lz
After an incubation period of 30 min at room temperature, MSLZ conjugate (25 µL, 3.3 g/L) and nephelometry buffer (125 µL) were added. All predilutions were performed in the nephelometry buffer with an automated dilutor (Hamilton). The scattered light was measured with the Sanofi-Diagnostics-Pasteur nephelometer Nephelia N600 (34) after incubation for 1 h at room temperature.
Reproducibility of the calibration curves was estimated by measuring light scattering for each dilution of the LZ calibrator in 10 successive assays. The precision of the immunoassay was assessed by measuring LZ in human milk samples with low, intermediate, and high concentrations 30 times during the same assay (within-run precision) and in 10 successive assays (between-run precision).
Analytical recovery was tested in a dilution-overloading experiment: the dilution assay was performed on two serial dilutions (1:1501:1200 and 1:2001:1600) of one milk sample containing 0.26 g/L of LZ; the overloading assay was performed in one milk sample (0.25 g/L) overloaded by eight increasing amounts of purified LZ (from 0.03 to 0.60 g/L). The slopes calculated by linear regression analysis for the dilution and the overloading assays were compared using Student's t-test.
For the total recovery, including dilution and overloading assays, the null hypothesis H0 (intercept = 0 and slope = 1) vs the alternative hypothesis H1 (intercept 0 and slope 1) were tested by F (Fisher) and t-tests, respectively, for intercept and slope.
Possible interference of -lactalbumin in LZ determination in human milk was assayed by overloading aliquots of one milk sample (0.33 g/L of LZ, 2.2 g/L of -lactalbumin) with four increasing amounts of purified -lactalbumin (from 1.1 to 4.9 g/L). Linear regression parameters of LZ recovery in these loaded samples were analyzed as above.
Application in other body fluids
The MSLZ conjugate remained immunoreactive for at least 6 months when stored at -20 °C. It was agglutinated by serial dilutions of anti-LZ As, and the light scattered by conjugate clusters formed during this immunoagglutination could be quantified by nephelometry.
Agglutination of MSLZ conjugate (275 mg/L) with anti-LZ As (900-fold diluted) was progressively inhibited by graded concentrations of free LZ (8500 µg/L in the reaction mixture). Fifty percent inhibition was observed with 67 µg/L LZ, and a minimal concentration of 8 µg/L was detectable in the reaction mixture as yielding an intensity of light scattered 3 SD lower than the mean value obtained in the absence of LZ (0% inhibition).
milk lz assay
A calibration range from 0.09 to 1.5 g/L of LZ in whole human milk (Fig. 1 ) was thus obtained when the assay was performed, as indicated in Materials and Methods, with milk samples diluted 6000-fold in the reaction mixture. Reproducibility CVs (n = 10) of the light scattering measured for each concentration of the LZ calibrator used for calibration were <4%.
The precision of the milk LZ immunoassay was assessed by the CVs obtained in within- and between-run studies (from 1% to 5%) as shown in Table 1 . Analytical recovery (Fig. 2 ) was linear (n = 16, r = 0.994, P <0.001) for the range of LZ concentrations in human milk tested in dilution-overloading assays (0.11.0 g/L). A mean percentage of recovery of 101% (SD, 10%) was obtained. The slopes of the dilution (0.932) and overloading (1.051) curves were not significantly different (P >0.05).
The slope (0.960) and the intercept (0.013 g/L) of the total recovery curve, including dilution and overloading assays, were not significantly different (P >0.05) from 1 and 0, respectively. Increasing amounts of -lactalbumin added in milk samples did not interfere with LZ assay (Fig. 2 ): a mean percentage of recovery of 103.3% (SD, 3.3%) was obtained for LZ measurement, and the slope (0.004) of the -lactalbumin overloading assay was not significantly different from 0 (P >0.05).
Application to human milk samples
Changes in milk LZ concentration and in the proportion of LZ in milk total proteins were observed according to the stage of lactation (Fig. 3 ). The global concentration per liter was significantly higher (P <0.05) in colostrum from days 15 (n = 168; mean, 0.36 g/L; SD, 0.27 g/L) than in transitional milk from days 614 (n = 182; mean, 0.30 g/L; SD, 0.19 g/L) and remained stable in the mature milk from days 1528 (n = 167; mean, 0.30 g/L; SD, 0.12 g/L) and days 2942 (n = 63; mean, 0.30 g/L; SD, 0.14 g/L). The LZ concentration then increased in the mature milk from days 4356 (n = 28; mean, 0.35 g/L; SD, 0.07 g/L) and especially (P <0.001) in the mature milk from days 5784 (n = 28; mean, 0.83 g/L; SD, 0.24 g/L). The proportion of LZ among total proteins was found to be always rising (P <0.05) during lactation from colostrum (mean, 1.7%; SD, 1.1%) to the mature milk from days 5784 (mean, 7.3%; SD, 1.9%).
Application in other body fluids
The concentration of LZ was found to be <0.08 mg/L in eight urine samples and ranged from 0.08 to 1.12 mg/L for the 42 other (mean, 0.35 mg/L).
In spite of the conservation of 40% of amino acid residues between human -lactalbumin and LZ (9), no interference was observed in a experiment in which human milk was overloaded with high amounts of human -lactalbumin.
The microparticle-enhanced nephelometric immunoassay of LZ thus developed in human milk is easy to perform, without washing or phase separation, and rapid (90 min, 240 results/h), allowing LZ measurement in large series of milk samples.
The concentration of LZ in milk can be measured over a large calibration range (from 0.09 to 1.50 g/L) with high reproducibility (CVs <5% within and between runs) and accuracy (linear recovery in dilution-overloading assay). The inhibition mode chosen protects against the risk of underestimation by antigen excess.
The lower detection limit (8 µg/L) allows to use high dilutions of milk samples, cancelling sample blank measurement and such clarifying pretreatment as skimming or casein precipitation.
In comparison, the mostly used methods for LZ determination present several limitations and drawbacks: a poor lower detection limit (immunoelectrophoresis, 50 mg/L (20), enzymatic determination, 15 mg/L (17)(18)(19), and conventional immunonephelometry, 1 mg/L (23)(24)); the need to pretreat samples (17)(19)(23)(24); the possible action of factors in biological fluids that alter the enzymatic activity of LZ (35); a long incubation period (18 h were necessary to obtain reliable results by the lysoplate technique ((19))); the constraints because of washing and phase separation (21) or the use of radioisotopes (22); and the risk of underestimation by antigen excess, which may be encountered in immunoassays based on a noncompetitive antigen-antibody reaction (21)(23)(24).
The concentrations of LZ measured by this microparticle-enhanced nephelometric immunoassay in 636 human milk samples, collected from 74 mothers and including colostrum and transitional and mature milks, were individually distributed over a large range (0.091.83 g/L) with mean concentrations ranging from 0.30 to 0.83 g/L in relation with the lactation stage. The comparison of these mean concentrations of LZ in human milk with those (0.021.5 g/L) previously reported (10)(11)(12)(13)(14)(15)(16) was difficult because of the lack of standard material. These last results were, incidentally, highly debated according to the methods used to obtain them, whether they were enzymatic or immunochemical determinations, and in terms of LZ standards and sample pretreatment (14)(22)(24).
The large distribution of milk LZ concentrations reflected both great individual variability and the influence of the stage of lactation. Variations in LZ concentrations of human milk, in relation with age, parity, maturity of pregnancy, and the mother's diet, have already been reported (12)(13)(14)(15)(16).
Important changes in the absolute LZ concentrations in human milk during lactation were also observed previously (10)(11)(12)(13)(14)(15). Our results, showing a nadir of LZ concentration (0.30 g/L) from 2 to 6 weeks of lactation and a progressive increase (0.83 g/L during the third month) are similar to these reported in other studies. The constant increase of the relative concentration of LZ among milk total proteins from colostrum (1.7%) to mature milk from days 4356 (3.8%) can be identified as being principally the consequence of the decrease (from 22 to 10 g/L) of the total protein content of human milk during the same period. Such decreases occur earlier for secretory IgA (14) and lactoferrin (29), and are more staggered for -lactalbumin (28) and ß-casein (30). The strong increase of the relative milk LZ concentration after the second month of lactation (7.3% in mature milk from days 5784), at a time when the total protein content of human milk slightly increases again (11 g/L), suggests that LZ, as well as lactoferrin (29), could play a major part as antiinfectious agents in the passive protection of breast-fed infants during mature lactation.
They could also be involved in the local protection of the mammary gland itself.
A microparticle-enhanced nephelometric immunoassay of LZ was applied here in a large series of human milk samples to study quantitative changes during lactation.
Because of its ease and rapidity, such an immunoassay could also be an appropriate method for allowing an exhaustive investigation of LZ milk changes in relation to the mother's status. The applicability of microparticle-enhanced nephelometric immunoassays in various body fluids has been reported previously (25)(26)(34)(36). The present application to the determination of LZ in serum and saliva, the results of which were close to those previously reported (3)(19)(20)(23)(24), suggests that microparticle-enhanced nephelometric immunoassay could be used more widely to quantify this non-antibody immune factor in other external secretions than milk and to investigate its biological importance. However, the lower detection limit will have to be improved by optimization of assay conditions to measure the lowest concentrations of LZ in urine (37)(38).
1 Nonstandard abbreviations: LZ, lysozyme; As, antiserum; and MS, microsphere.