Main

Platelet counts in newborns are similar to those of adults or children(1–3). However, newborn infants admitted to intensive care nurseries have a high prevalence of thrombocytopenia(4–7). The mechanism(s) responsible for this increased susceptibility to thrombocytopenia is unclear(7). In addition to thrombocytopenia, some studies have documented functional abnormalities in newborn platelets(8–10), whereas others have not(11). During maturation, Mk undergo endoreduplication yielding a polyploid cell. Maturational differences in fetal Mk may be related to both quantitative and qualitative platelet differences observed in newborns. We have previously shown that fetal Mk size is smaller than adult Mk size, a parameter that may also correlate with cellular DNA content(12).

Several techniques to measure Mk ploidy are currently available(13–16). The most frequently used techniques involve cell sorting(16). We previously described a modified Feulgen technique using bone marrow aspirate fixed smears to study ploidy(15). When studying fetal samples, a large quantity of bone marrow is not available. Faced with small fixed tissue samples, we developed a technique for estimation of DNA content using biopsy specimens by a modification of our previously described technique(15). Mk DNA content measurement was performed on samples from seven adult volunteers, comparing results between bone marrow biopsies and bone marrow aspirate specimens to confirm the validity of the method.

In an effort to understand differences between platelets in newborns and in adults, we examined Mk ploidy in the fetus and compared it with adult bone marrow, using the modified Feulgen stain to measure DNA of individual Mk from small biopsy specimens.

METHODS

Human subjects. Bone marrow aspirates and trephine biopsies from the iliac crest were collected from seven healthy adult volunteers after obtaining informed consent. Human fetal tissue was obtained from 15 normal and after vacuum aspiration from 2 anencephalic abortuses of 12-21-wk gestation. Biopsy specimens of bone marrow were obtained from femurs. All studies were approved by the Institutional Review Board and carried out according to the principles of the Declaration of Helsinki.

Bone marrow smears. Bone marrow aspirates were collected in clear syringes and immediately prepared on glass slides, allowed to air dry, fixed in methanol:acetone 1:1 for 10 min, rinsed in water, and then stored frozen at -20°C. To prevent the specimen from washing away during the staining process, glass slides were coated with 0.1% poly-L-lysine and allowed to dry in a dust-free chamber.

Biopsy preparation. Biopsy specimens were fixed in 3% paraformaldehyde for 4-8 h, serially dehydrated in acetone, and then embedded in glycolmethacrylate (JB4; Polysciences, Inc., Warrington, PA). Sections were cut 3 μm thick with glass knives and mounted on glass slides. To prevent the specimen from washing away during the staining process, glass slides were coated with 0.1% poly-L-lysine and allowed to dry in a dust-free chamber(17).

Fluorescent staining. A modified Feulgen technique was adapted and developed from the method of Mazur et al.(15). The compound BAO (Fluka Chemical Corp., Buchs, Switzerland) is relatively stable to UV light. BAO has been recommended for human bone marrow Mk; such staining is comparable to auramine O in its specificity for DNA and has been shown to be approximately 96% DNA-specific(15). The specimens were hydrated in 0.1 M PBS for 20 min. DNA was hydrolyzed in 4 N HCl at 37°C for 30 min, then washed in distilled water at room temperature to terminate hydrolysis. BAO was prepared with 20 parts of 0.01% BAO plus 2 parts of 1 N HCl plus 1 part of 10% sodium(meta) bisulfite. The slides were stained with fresh BAO with agitation in a covered incubation chamber (no light) at room temperature, 30-45 min. This was followed by three 2-min washes in sulfite water (1 part of 10% bisulfite plus 1 part of 1 N HCl plus 18 parts of distilled water), then washed in distilled water and allowed to dry in the dark.

DNA quantitation. An Olympus BH-2 microscope equipped with a fluorescent lamp and interchangeable aperture to a Zeiss photometer was used to measure BAO fluorescent signal photometrically. Aperture size was 0.6 mm for biopsy specimens; 1.0 mm was needed for bone marrow smears due to cell spreading on the slides.

At 100×, Mk were identified morphologically and centered in the aperture. With UV light excitation at 400×, the BAO fluorescent signal was measured. A granulocyte nucleus in the same or adjacent field was also measured as the diploid reference. Background readings were recorded in each field. Although the specimen remained quite stable, care was taken to limit the ambient light and UV excitation on the specimen.

Maturational stage. After recording the light emission, the maturational stage of each Mk was assessed as stage II, III, or IV, guided by morphologic criteria of Levine et al.(18). Stage I Mk, the least mature cell or megakaryoblast, cannot be reliably distinguished from other blast cells in the marrow by morphology alone, and therefore was not recorded. Stage II cells have larger nuclei with early nuclear lobulation and a comparatively high nuclear:cytoplasmic ratio. Stage III Mk are larger cells with abundant cytoplasm and complex multilobulated nuclei and separated nuclear lobes. Stage IV cells are very large with abundant cytoplasm and compact multilobulated nuclei.

Calculation of DNA content. The relative DNA content of each Mk, based on the diploid (2N) granulocyte, was calculated by first subtracting the background measurement from each cellular reading, then dividing the net Mk DNA emission signal by the net diploid cell measurement. The result was multiplied by two to be in conventional ploidy units. For each cell:Equation

Statistical analysis. The arithmetic and geometric means and standard deviations were calculated as for grouped data. To calculate the geometric mean, the log10 of each individual ploidy measurement was determined, the arithmetic mean of the log10 ploidy values for each specimen was determined, then the antilog10 of that arithmetic mean was calculated to return to units of ploidy: Equation wheren is the number of Mk examined. Means were compared using the two sample t test for samples with equal or unequal variance (confirmed by F test). Mk ploidy and stage distributions were compared using Pearson χ2 test.

RESULTS

Smears and biopsy specimens from seven healthy young adults were examined and compared. After initial trials, the incubation times for acid hydrolysis and BAO staining of the smears were reduced to decrease difficulty with bone marrow cells loosening and washing away. Under usual conditions, it took approximately 1-2 h to examine one biopsy section and 4-6 h to examine a bone marrow smear. Mk were easier to find and more plentiful on biopsies. The mean number of Mk examined on the smears was 29, with a range of 17 to 54. The mean number examined on biopsy sections was 45, with a range of 31 to 55.

In each specimen, a continuous distribution was observed in both the whole cell and biopsy specimens, as shown in Figures 1 and 2. Both the arithmetic and geometric means were calculated for each specimen. The arithmetic mean was used for comparison, whereas the geometric mean was used to correct the otherwise disproportionately weighted higher ploidy values in the expected distribution (2N, 4N, 8N, 16N, 32N, 64N, 132N).

Figure 1
figure 1

Normal adult: smear of whole cells, ploidy vs stage distribution.

Full size image
Figure 2
figure 2

Normal adult: bone marrow biopsy, ploidy vs stage distribution.

Full size image

There was no significant difference in results between the two methods for determining Mk ploidy. The mean ploidy derived from the whole cell preparations was 15.0N and on biopsy sections was 16.7N (p = 0.2260). The geometric means yielded smaller values, but similarly showed no difference between whole cells, 12.2N, and biopsies, 14.0N (p = 0.1062). See Table 1.

Table 1 Ploidy values: smear of whole cells vs biopsy section, arithmetic and geometric means
Full size table

For analysis of ploidy distribution, values were assigned to the nearest ploidy level(15). The mode and median ploidy value was 16N for both methods, with 8N and 32N as the second and third most frequently observed values (Table 2). The distribution by stage was similar for both methods, with the majority of cells being mature stage IV cells (Table 3). The mean ploidy increased with increasing stage; although there was no significant difference, measurements in the biopsy specimens tended to be greater than values from the smear specimens (Table 4).

Table 2 Frequency by ploidy distribution: smear of whole cells vs biopsy
Full size table
Table 3 Mk distribution by stage: number of cells examined in adult smear and adult and fetal biopsy specimens
Full size table
Table 4 Mean ploidy by stage: smear of whole cells vs biopsy and adult biopsy vs fetal biopsy
Full size table

The mean number of Mk examined for adult biopsy sections was 45 (range 31 to 55), and 45 for fetal biopsy sections (range 6 to 91). As previously described, fetal Mk showed a shift to the earlier stage, immature Mk(Table 3)(12).

Ploidy measurements showed a continuous distribution(Fig. 3). The overall mean ploidy of fetal Mk was decreased compared with adults. The difference was significant whether a geometric or an arithmetic mean was used. The geometric mean was calculated to correct the otherwise disproportionately weighted higher ploidy values in the expected distribution (2N, 4N, 16N, 32N, 64N, 128N,...). The arithmetic mean is shown in Table 4 for both fetal and adult Mk ploidy. Similar results were obtained with geometric means of 7.85 for fetuses and 13.0 for adults (data not shown).

Figure 3
figure 3

Combined ploidy distribution from 10 fetal samples.

Full size image

Differences were most evident in the more mature stages III and IV Mk. Although fetal Mk DNA content increased only slightly as maturational stage increased, the mean adult Mk DNA content doubled from stage II to stage IV(see Table 4). Over the gestational period studied, there was no statistically significant relationship between gestational age and Mk DNA content.

DISCUSSION

Susceptibility of newborns to thrombocytopenia differs from older children and adults, suggesting that the precursor cell, the Mk, differs in fetuses compared with adults(4–7). In an attempt to understand this difference we study Mk ploidy as a parameter of Mk maturation. This study utilizes a method for estimation of Mk ploidy distribution developed to deal with the circumstances where an appropriately enriched marrow cell suspension is not available for flow cytometry or for smear on glass slides.

The method of examining cut cells on biopsy sections yields results which are not different from results using whole cells. An entire cross section of a trephine biopsy can be examined more quickly and thoroughly than smears of whole cells on glass slides, without the difficulty of losing the specimen during processing. As expected, the geometric mean yielded a smaller value for the mean ploidy, but with no significant difference in the closest modal value of 16N. This method did not include stage I or all of the stage II Mk, which are morphologically indistinguishable from other blasts in the marrow, thus incompletely representing the 2N, 4N, and 8N populations. The group of immature Mk precursors have been identified in other studies; using monoclonal or polyclonal antibodies to surface glycoproteins(19).

Previous studies of normal human marrow have shown similar results, finding 16N as the modal ploidy, with 8N and 32N as the next most frequent classes(18, 20). Results using the biopsy method are similar to results of previous investigators using conventional methods, flow cytometry, and Feulgen techniques on whole cells. Two other investigators have used fixed tissue samples and found similar results(13, 21). This biopsy method should be recognized as providing only an estimate of ploidy, because whole cells cannot be assured in cut sections.

Until now, there has been limited availability of a method of determining ploidy under conditions where an aspirate of adequate cell suspension is unavailable, such as myelofibrosis or in infants and fetuses. Additionally, the biopsy method offers the advantage of showing a representative population, unaltered by the process of collection, isolation, and purification. This method of determining ploidy distribution from bone marrow biopsy specimens gives a reliable estimate of Mk ploidy and should be considered when conventional preparations of whole cells cannot be obtained.

Applying this method to fetal bone marrow biopsies, we demonstrated that the overall mean ploidy of fetal Mk is decreased compared with adults. Additionally, fetal Mk ploidy increases as maturational stage increases, but not to the same extent that adult Mk ploidy increases with Mk maturation. Over the gestational period studied, there was no relationship between gestational age and mean ploidy.

This method did not include the most immature Mk. Using an immunoperoxidase-linked stain previously described(12) we attempted to identify these cells, but the combined immunoperoxidase stain and BAO stain did not produce reproducible data (not shown). Because the majority of these immature Mk are low ploidy, their exclusion should not affect the findings of this work.

Previously we showed that fetal Mk are smaller and less mature than adult Mk(12). Olson et al.(22) have shown that cord blood contains an increased number of Mk compared with adult peripheral blood. The ploidy of these Mk is also decreased compared with adults(22). In anin vitro culture system, using Mk from fetal liver, cord blood, and adult bone marrow, Hegyi et al.(23) showed that cultured fetal Mk are smaller and have a lower ploidy than adult Mk. These investigators suggested that a confirmatory test looking at in vivo Mk is necessary because cultured Mk tend to be smaller and less polyploid that those seen in vivo(23). Increase in Mk ploidy is an important aspect of the normal response to thrombocytopenia(24, 25). The decreased ploidy of Mk in the fetus may impair this response. In vitro fetal Mk increase their ploidy on stimulation by IL-6(26). However, there is no information on the ability of fetal or newborn Mk to respond with polyploidization to the stress of thrombocytopenia. The small size, shift to a less mature population and decreased ploidy of fetal Mk indicate that there are differences in the postmitotic phase of Mk development in the fetus. Such differences are likely related to quantitative and qualitative platelet abnormalities in the newborn. Understanding the physiology and regulation of megakaryocytopoiesis in the fetus and newborn will be valuable in determining the pathophysiologic basis of platelet dysfunction in the newborn.