Main

PSA-NCAM is a developmentally regulated posttranslational modification of NCAM (1). Due to steric properties, the highly negative charge and the heavily hydrated molecule, PSA is known to interfere with the adhesive functions of NCAM and other cell surface molecules, thereby modulating various cell-cell interactions (2,3). PSA-NCAM is abundant in embryonic tissues, decreases progressively during development, and becomes restricted to regions of persistent proliferation and neurogenesis in the adult (4,5). Several malignant tumors, such as small cell lung cancer (6,7), multiple myeloma (8), neuroblastoma (9), and rhabdomyosarcoma (10), reexpress PSA-NCAM on the surface of their tumor cells. Recently, patients suffering from these tumors were shown to have distinctly elevated PSA-NCAM serum levels, proving this molecule to be a valuable diagnostic and prognostic tumor marker (11–17).

In healthy adult blood donors, serum PSA-NCAM concentrations usually are below 20 kU/L, and therefore this value was defined as the upper limit of the normal range (13,15,16). Our previous study (12) revealed that in healthy children the levels are on average higher compared with normal adults. In this study, by assigning children of different ages to half overlapping age groups and analyzing their serum concentrations, we establish an age-dependent decrease of the serum levels and describe their statistical characteristics. For the first time, we determine reference ranges of serum PSA-NCAM for normal children of different age.

METHODS

Patients. Samples of blood (1.5 mL or less) were taken from 324 otherwise healthy, mostly German children, aged 1 d to 17 y, who were admitted for elective surgery, such as inguinal hernia, circumcision, or cryptorchidism, or for diagnostic measures for minor pediatric surgical diseases. The sex ratio was male:female = 206:118 (64% male predominance). Preterm babies were excluded, and none of the children had clinical evidence of internal disease nor significant medication. An oral consent was given by the parents in each case, and the study was approved by the local ethical and scientific committees. Another 42 samples were obtained from venous cord blood of spontaneously at term delivered, healthy neonates immediately after birth. A third group of patients consisted of 18 adult individuals, aged 22-84 y. No one of these individuals had the diagnosis of a malignant disease at the time of blood sampling. Directly after blood withdrawal, serum was obtained, frozen, and stored at -80°C until measured. We have published previously data of part of the patients (12).

Assay. Serum samples were thawed and then assayed by a chemiluminescent immunoassay based on coated tubes, as described previously (11,18). In this assay system, two MAb were used: MAb 735, which specifically detects and binds polysialic acid (19) as the capture antibody, and the anti-human NCAM MAb BW SCLC-1, conjugated to an acridinium N-acylsulfide label for detection of the bound PSA-NCAM. The results were expressed as kU/L. All measures were done at least in duplicate for internal control.

Assay characteristics. Analytical sensitivity was <0.5 kU/L; linearity and recovery study results were within 10% of the expected concentration; precision (10 samples, 8 measurements per sample) was 3.1-11.2% CV, mean 7.4%; and interassay precision (8 samples, 1-wk period, 2 measurements a day) was 4.9-8.7% CV, mean 6.1%. Hyperbilirubinemia, hyperlipidemia, and hemolysis had no influence on the precision of the assay system.

Statistics. Data analysis and statistical tests were performed using SigmaPlot version 2.01, Jandel Corporation, Microsoft Excel version 5.0, and SPSS, on a PC-Pentium166, 20MB RAM, 2GB HDD, under Microsoft Windows 95.

RESULTS

Age groups. The patients were divided into half-overlapping subgroups of 1000 d, newborn sera (NS = cord blood), and adult patients, which led to 14 groups containing at least 10 members. Table 1 shows the parameters for all groups: age, number of members, mean serum concentration, SD, and empirical interpolated percentiles at 1, 5, 50 (= median), 95, and 99%. In Figure 1 the decrease of mean PSA-NCAM concentration and standard deviations are shown as a function of age. The highest serum levels were found in the sera of cord blood (51.7 ± 12.0 kU/L, mean ± SD), whereas the lowest values were found in the age group of adult patients (9.9 ± 3.5 kU/L, mean ± SD). In between, a defined decay of PSA-NCAM serum concentrations in dependence on age was found (see below).

Table 1 Age groups, number of patients, mean, SD, and percentiles for serum PSA-NCAM levels in normal children
Full size table
Figure 1
figure 1

Mean PSA-NCAM serum concentrations (bold line) and standard deviations (thin lines) in dependence of age.

Full size image

Male-female difference. Subgroups included male and female patients. To test the null hypothesis that NCAM values of male and female patients were identical within age groups, one t test per age group was performed. Table 2 shows age ranges, number of male to female patients, mean and SD of each group, and test results for double-sided t tests. On a significance level of p < 0.05 the null hypothesis could not be rejected. We accepted the null hypothesis for all age groups, and therefore, we did not further distinguish between male and female.

Table 2 Double-sided t test on each age group to test the difference between male and female patients, which was not statistically significant (p < 0.05)
Full size table

Trend detection. To detect a trend between two subgroups, neighbored groups with nonoverlapping age ranges were tested against the null hypothesis of both groups being identical. Tests were performed with t tests. Table 3 shows both age ranges, number of members, and test results for the double test. On a significance level of p < 0.001, the null hypothesis had to be rejected for newborn sera against the first age group, a highly significant difference. The trend toward adults allowed us to reject the null hypothesis on a level of p < 0.001 as well, which we got for the single-sided alternative hypothesis between the latest tested age groups.

Table 3 Double-sided t test on neighboured nonoverlapping age groups to detect a trend between the age groups
Full size table

Fitted logistic function. To quantify the decay of PSA-NCAM serum concentrations from small infants toward adult patients, we fitted a logistic function and, therefore, produced some remarkably handy parameters. Fitting was performed with SigmaPlot. To describe the dynamics during the entire childhood, we excluded the newborn sera. Figure 2 shows the type of logistic function, fitted parameters, and their SEM. a ± SE = 29.5 ± 0.7 kU/L represents the mean initial PSA-NCAM value up to approximately 5 y of age, d ± SE = 9.3 ± 1.8 kU/L the mean PSA-NCAM value of adult patients, and most interesting c ± SE = 5144 ± 353 d ≈ 14 ± 1 y the half-time of the decrease, demonstrating that PSA-NCAM serum concentrations in average reach half of the levels in small infants at an age of 14 ± 1 y.

Figure 2
figure 2

Fitted logistic function to describe the age-dependent decay of PSA-NCAM serum concentrations during childhood. Cord blood samples were omitted from fitting (see text).

Fitted logistic function: f(age) = (a - d)/(1 + (age/c)⁁b) + d Table

Full size image

The same type of logistic function was fitted to PSA-NCAM serum concentrations of the children, including cord blood samples. By using a semilogarithmic diagram and thereby closely looking at the newborn period, another sigmoid graph was obtained (Fig. 3). Half-time of decay was determined at c ± SE = 3.1 ± 2 d; mean initial PSA-NCAM was a ± SE = 52.0 ± 1.5 kU/L. Childhood fitted PSA-NCAM levels d ± SE = 27.9 ± 0.8 kU/L correspond well to the values obtained by dividing the children into overlapping age groups as displayed in Table 1.

Figure 3
figure 3

Fitted logistic function of PSA-NCAM serum concentrations from children age up to 1000 d (about 2½ y) including cord blood samples. The semilogarithmic diagram closely shows the newborn period and early infancy.

Fitted logistic function: f(age) = (a - d)/(1 + (age/c)⁁b) + d Table

Full size image

The coefficient of correlation was calculated according to Spearman for nonlinear relations and was found in both cases to be 0.33. Because of the wide spreading data the correlation appears weak, but fitting the sigmoid functions with SigmaPlot still allows us to derive some half-times of decay with acceptable SEM.

Follow-up measurements. In 34 patients, follow-up measurements were done on different days at intervals from 1 d to 7 mo. Showing variations below 10% (mean SD 2.6 kU/L, range 0.1-9.3 kU/L), follow-up values were not different from the initial value.

DISCUSSION

The highest PSA-NCAM serum levels (51.7 ± 12.0 kU/L, mean ± SD) were found in cord blood samples, consisting almost exclusively of fetal blood from term delivered (40 ± 2 gestational wk) healthy newborns. In contrast to the relatively narrow time range of this group, the next group covered children aged 1-1000 d (= 2 y, 9 mo). The wide age range is possibly responsible for the abrupt decrease of the serum levels to a mean value of 30.0 (± 10.5) kU/L in the second group. Thereafter, a continuous decrease of the levels was observed until adolescent age (Table 3). The transition into the adult age group was characterized by another relatively abrupt decrease of the serum concentrations, possibly caused by the wide age range of adult patients, who were defined as patients older than 17 y. In accordance with the literature (13,15,16), the levels in this group consistently were below 20 kU/L (9.9 ± 3.5 kU/L, mean ± SD). The organization of the children into overlapping age groups produced very handy parameters despite scattering of the levels over a wide range, as seen in Table 1 and Figure 1. Subdividing the age groups into smaller sections or considering nonoverlapping age groups did not improve the results.

The decrease of the PSA-NCAM serum levels during the newborn period and during childhood were described by specially fitted logistic functions (Fig. 2 and 3). Obviously, there is a first decay with a half-time of 3.1 ± 2 d (mean ± SE), and a second decay covering childhood with a half-time of 14 ± 1 y (mean ± SE). Maybe there are main PSA-NCAM-dependent developmental processes abruptly ending at birth, whereas others last until young adulthood, explaining the persistent elevation of the serum levels during childhood.

Parameters such as total body mass, body surface, head circumference, or weight of the patients had no effect on PSA-NCAM serum concentrations in the children tested. Because the parameters were not consistently specified, they are not further mentioned herein. The study was performed in mostly German children without anamnestic evidence of disease, undergoing minor pediatric surgical operations, and therefore, no other serum measures were determined. As expected, we found no statistically significant difference of the serum levels between male and female patients.

The mechanisms how NCAM forms appear in the serum are not exactly known. Likely, there is a constant release of polysialylated NCAM from cell surfaces, and serum levels of an individual patient reflect the amount of PSA-NCAM positive cells of the body (11). Especially in newborns and small infants, we found the levels covering wide ranges, reflected by high standard deviations of the corresponding age groups. In contrast, repeated measurements in individual patients produced only little deviations from the expected values, and standard deviations decreased continuously until adulthood (from 12.0 kU/L in cord blood samples to 3.5 kU/L in samples from adult patients). Possibly, this reflects the biologic variation of PSA-NCAM-dependent developmental processes in different patients ending at a common level in adulthood. It is not known where and how fast serum NCAM forms are degraded. At least, they are-as expected by the steric properties and highly negative charge of the large molecule-not excreted in the urine (our unpublished observations).

In conclusions, we present age-dependent reference values and empirical percentiles for serum PSA-NCAM in normal children. On the assumption that serum concentrations reflect the amount of PSA-NCAM-positive cells of the body, our data indicate that developmental processes dependent on this molecule last far beyond the perinatal period until they come to an end in adulthood.

Table 4
Full size table
Table 5
Full size table