Rat spermatogenesis in vitro traced by quantitative flow cytometry

In vitro differentiation of germ cells in rat seminiferous tubule segments at stages II-III of the epithelial cycle was studied. DNA flow cytometry was used for quantitation of absolute cell numbers from the cultured tubule segments that were compared to freshly isolated stages of the cycle, as identified by transillumination stereomicroscopy of the seminiferous tubules and phase-contrast microscopy of live cell squashes. Spermatogonia and spermatocytes from stages II-III showed normal morphological differentiation during 7 days in vitro. Round spermatids differentiated to Step 7 of spermiogenesis but Step 16 spermatids failed to develop. Acid phosphatase activity in the spermatogenic cells changed normally during the culture. As compared with freshly isolated control tubule segments, 35% of round spermatids and 42% of pachytene spermatocytes were present in culture after 7 days. The cell numbers recovered from defined stages by DNA flow cytometry were close to those found in morphometric studies. Flow cytometry is an efficient quantitation method for cells liberated from seminiferous epithelium. Spermatogonia, spermatocytes, and early spermatids are able to differentiate in vitro, but spermatids approaching the elongation (acrosome) phase, and particularly the maturation phase, fail to differentiate under present culture conditions.     

Morphology of the Nephron in the Mesonephros of Bufo bufo (Amphibia,Anura, Bufonidae)

The present study deals with the morphology and ultrastruclure of the nephron in the mesonephros of the toad, Bufo bufo (Linnaeus, 1758). Based on serial sections in paraffin, Araldite and Epon, the position of the different segments of the nephron within the kidney tissue was determined, and a nephron subsequently reconstructed. The nephron consists of the following parts: Malpighian corpuscle, neck segment, proximal tubule, intermediate segment, early distal tubule, late distal tubule and collecting tubule. The late distal tubule was subdivided into three morphologically different sections. The total number of nephrons in the toad mesonephros was estimated at 6000 units. The length of the segments in the reconstructed nephron was calculated. The cytology of the epithelial cells constituting the segments was described using transmission and scanning electron microscopy. Heterocellularity was found in the late distal tubule section I and III and in the collecting tubule. The proportional distribution and number of intercalated (mitochondria-rich) cells in the late distal tubule and collecting tubule was calculated. Only one morphological type of intercalated cell could be distinguished. Late distal tubules were removed from fresh Bufo kidneys for preliminary studies of the intercalated cells with Nomarski optics.     

Effect of vascular endothelial growth factor and testis tissue culture on spermatogenesis in bovine ectopic testis tissue xenografts

Bovine ectopic testis tissue grafting is a technique that can be used to study bovine spermatogenesis and for the production of germ cells for a variety of applications. Approximately 10% of seminiferous tubule cross sections in testis grafts contain spermatids, providing a unique tool to investigate what regulates germ cell differentiation. We hypothesized that manipulation of testis tissue grafts would increase the percentage of seminiferous tubule cross sections undergoing complete germ cell differentiation. To test this hypothesis, bovine testis tissue was treated with vascular endothelial growth factor (VEGF) at the time of grafting or explant cultured for 1 wk prior to grafting. For the VEGF experiment, 8-wk donor tissue and graft sites were treated with 1 microg of VEGF in order to increase angiogenesis at the graft site. For the testis tissue culture experiment, 4-wk-old donor testis was cultured for 1 wk prior to grafting to stimulate spermatogonial stem cell proliferation. Testis tissue grafts were removed from the mice 24 wk after grafting. VEGF treatment increased graft weight and the percentage of seminiferous tubule cross sections with elongating spermatids at the time of graft removal. Cultured testis tissue grafts were smaller and had fewer seminiferous tubules per graft. However, there was no difference in the percentage of seminiferous tubule cross sections that contained any germ cell type between groups. These data indicate for the first time that bovine testis tissue can be manipulated to better support germ cell differentiation in grafted tissue.     

Cyclic expression of endothelin-converting enzyme-1 mediates the functional regulation of seminiferous tubule contraction.

The potent smooth muscle agonist endothelin-1 (ET-1) is involved in the local control of seminiferous tubule contractility, which results in the forward propulsion of tubular fluid and spermatozoa, through its action on peritubular myoid cells. ET-1, known to be produced in the seminiferous epithelium by Sertoli cells, is derived from the inactive intermediate big endothelin-1 (big ET-1) through a specific cleavage operated by the endothelin-converting enzyme (ECE), a membrane-bound metalloprotease with ectoenzymatic activity. The data presented suggest that the timing of seminiferous tubule contractility is controlled locally by the cyclic interplay between different cell types. We have studied the expression of ECE by Sertoli cells and used myoid cell cultures and seminiferous tubule explants to monitor the biological activity of the enzymatic reaction product. Northern blot analysis showed that ECE-1 (and not ECE-2) is specifically expressed in Sertoli cells; competitive enzyme immunoassay of ET production showed that Sertoli cell monolayers are capable of cleaving big ET-1, an activity inhibited by the ECE inhibitor phosphoramidon. Microfluorimetric analysis of intracellular calcium mobilization in single cells showed that myoid cells do not respond to big endothelin, nor to Sertoli cell plain medium, but to the medium conditioned by Sertoli cells in the presence of big ET-1, resulting in cell contraction and desensitization to further ET-1 stimulation; in situ hybridization analysis shows regional differences in ECE expression, suggesting that pulsatile production of endothelin by Sertoli cells (at specific "stages" of the seminiferous epithelium) may regulate the cyclicity of tubular contraction; when viewed in a scanning electron microscope, segments of seminiferous tubules containing the specific stages characterized by high expression of ECE were observed to contract in response to big ET-1, whereas stages with low ECE expression remained virtually unaffected. These data indicate that endothelin-mediated spatiotemporal control of rhythmic tubular contractility might be operated by Sertoli cells through the cyclic expression of ECE-1, which is, in turn, dependent upon the timing of spermatogenesis.     

Localization and synthesis of entactin in seminiferous tubules of the mouse.

Localization and synthesis of entactin in seminiferous tubules of mouse testis was studied by immunocytochemistry. Frozen sections from adult mice testes were subjected to anti-entactin and anti-laminin immunofluorescence. Both entactin and laminin were localized within the seminiferous tubule basement membrane and intertubular region of the testis. The addition of excess amount of entactin (but not fibronectin), premixed with anti-entactin antiserum, abolished the immunostain. Western blotting showed that a protein extract from a seminiferous tubule basement membrane preparation was recognized by anti-entactin anti-serum and comigrated with recombinant entactin. Enriched fractions of isolated primary Sertoli cells and peritubular myoid cells cultured for 6 days on a glass coverslip were able to synthesize and secrete entactin as detected by immunofluorescence microscopy. Entactin was also produced by TM3 (Leydig-like) and TM4 (Sertoli-like) cell lines as detected by both immunofluorescence and Western blotting. The distribution of entactin vs. laminin within both the cultured primary cells and the TM3 and TM4 cell lines differed. Entactin appeared mainly localized extracellularly. In contrast, laminin was mainly localized intracellularly. The above findings suggested that 1) entactin existed in the seminiferous tubule basement membrane and intertubular region of adult mice testis, co-localized with laminin; 2) entactin was synthesized by the cultured primary Sertoli cells and peritubular myoid cells and the TM3 and TM4 cell lines; 3) entactin was exocytosed with little intracellular accumulation, in contrast to an intracellular accumulation of laminin.     

Centromere pattern in different mouse seminiferous tubule cells

Centromere arrangement in different mouse seminiferous tubule cells was analyzed using an anticentromere antiserum from a patient with the CREST syndrome of scleroderma. A peptide of 18 kd was recognized by this serum on immunoblotting of mouse nuclear proteins from seminiferous tubule cells. In the cells studied by immunofluorescence, different patterns of centromere arrangement were observed. A speckled arrangement of centromeres was found in spermatogonia, double spots corresponding to meiotic bivalents were found in pachytene cells, and clusters of a haploid numer of centromeres were found in early and acrosome phase spermatids. In Sertoli cells, only three centromeric spots were detected, corresponding to the nucleolar organizer chromosome pairs. A relationship between the functional stage of the cell and the arrangement of and conformational changes in the centromeres is considered.     

A comparative study in twelve mammalian species of volume densities, volumes, and numerical densities of selected testis components, emphasizing those related to the Sertoli cell

Morphometric studies were performed on 12 mammalian species (degu, dog, guinea pig, hamster, human, monkey, mouse, opossum, rabbit, rat, stallion, and woodchuck) to determine volume density percentage (Vv%), volume (V), and numerical density (Nv) of seminiferous tubule components, especially those related to the Sertoli cell, and to make species comparisons. For most species, measurements were taken both from stages where elongate spermatids were deeply embedded within the Sertoli cell and from stages near sperm release where elongate spermatids were in shallow crypts within the Sertoli cell. Montages, prepared from electron micrographs, were used to determine Vv% of Sertoli cell components in seminiferous tubules. Excluding the tubular lumen, the Sertoli cell occupied from a high of 43.1% (woodchuck) to a low of 14.0% (mouse) of the tubular epithelium. There was a strong negative correlation (r = -0.83; P less than 0.005) of volume occupancy of Sertoli cells with sperm production. Nuclear volume, as determined by serial reconstruction using serial thick sections, ranged from a high of 848.4 microns 3 (opossum) to a low of 273.8 microns 3 (degu). There was no correlation (r = 0.02) of nuclear volume with volume occupancy (Vv%) in the tubule. Sertoli cell volume was determined by point-counting morphometry at the electron-microscope level as the product of the nuclear size and points determined over the entire cell divided by points over the nucleus. Sertoli cell V ranged from 2,035.3 microns 3 (degu) to 7,011.6 microns 3 (opossum) and was highly correlated (r = 0.85; P less than 0.001) with nuclear size. However, there was no significant correlation between the Sertoli cell size (V) and volume occupancy (Vv%; r = 0.13) or sperm production (r = -0.21). Stereological estimates of the numerical density (Nv) of Sertoli cells ranged from a high of 101.9 x 10(6) (monkey) to a low of 24.9 x 10(6) (rabbit) cells per cm3 of testicular tissue. There was no correlation of numerical density of Sertoli cells with sperm production (r = 0.002). A negative correlation was, however, observed between the numerical density of the Sertoli cells and the Sertoli cell size (r = -0.79; P less than 0.002). Data from the present study are compared with those previously published. This is the first study to compare Sertoli cell morphological measurements using unbiased sampling techniques. Morphometric data are provided which will serve as a basis for other morphometric studies.     

Cell-substratum and cell-cell interactions promote testicular peritubular myoid cell histotypic expression in vitro

Peritubular cells, prepared from seminiferous tubules from testes of 20-day-old-rats, were seeded onto different substrata and cultured under varying conditions. When plated onto polystyrene or glass surfaces, peritubular cells assumed a typical fibroblast-like cell shape and cell association pattern, together with a fibroblast-like migration behavior. They maintained high rates of proliferation even after achieving confluency. In contrast, when peritubular cells were plated onto a seminiferous tubule biomatrix (ST-biomatrix) surface, they spread to form a continuous cell layer having a myoepithelioid histotype similar to that of peritubular myoid cells in the intact seminiferous tubule. The characteristics of the myoepithelioid histotype described include a squamous, polyhedral cell shape; a cobblestone-like cell association pattern, with closely apposing or slightly overlapping cell borders, and a very low mitotic index. When peritubular cells were plated onto laminin, collagen, fibronectin, heparin, or a liver biomatrix, a fibroblast-like pattern resulted, indicating that ECM components listed and liver biomatrix are unable to substitute for ST-biomatrix in maintaining normal myoepithelioid characteristics in vitro. In cocultures of Sertoli cells plated on top of peritubular cells, the peritubular cells directly in contact with Sertoli cell aggregates developed a myoepithelioid histotype, whereas peritubular cells in regions not in direct contact had a fibroblast-like histotype. The data are discussed in relation to the possible role of cell-cell interactions, and cell-substratum interactions, in the acquisition and stabilization of the histotype of peritubular cells in the seminiferous tubule during development.     

A cytological and cytoskeletal comparison of Sertoli cells without germ cell and those with germ cells using the W/WV mutant mouse.

The distribution of F-actin and intermediate filaments in the W/WV mouse was investigated by light and transmission electron microscopy, and fluorescence methods. No spermatogenic cells were detected in the seminiferous epithelium of the W/WV mouse. Its seminiferous tubule was one-half the diameter of that in the normal (+/+) mouse. The Sertoli cell which was an only component of the W/WV mouse seminiferous epithelium was decreased in height, but still retained the polarity as evidenced by light microscopy. The Sertoli cell organelles were similar in appearance when normal and mutant mice were compared. F-actin was recognized at ectoplasmic specialization (ES) of the W/WV mouse Sertoli cell and appeared similar to the normal mouse. However, the junction with ES was more extensive compared with that of the normal mouse Vimentin in the W/WV mouse Sertoli cell was distributed around the nucleus and extended towards the tubular lumen similar to the normal mouse. Its extension within the Sertoli cell trunk, however, was restricted to a lesser degree as compared with that in the normal. Thus, the subcellular Sertoli cell and the distribution of F-actin and intermediate filaments (vimentin) in the W/WV mouse Sertoli cell seemed not to be strikingly affected by lack of spermatogenic cells, suggesting minimal influence of germ cells on Sertoli cell cytology and cytoskeleton.     

Electron microscope analysis of tissue components identified and located by computer-assisted 3-D reconstructions: ultrastructural segmentation of the developing human proximal tubule

A method is described for ultrastructural analysis of renal tubules after precise identification of tubule segments by computerized 3-D reconstruction at the light microscope level. Semithin serial sections were cut of entire nephrons and 3-D coordinate information was obtained by digitization of tubule cross sections in the semithin sections. With the aid of the computer the tubule axis was traced from one section to the other. Precise lengths and positions of the tubules in three dimensions were calculated and stereoscopic images generated. The method was used to analyze the 3-D structure of developing human nephrons, and the ultrastructural development of the proximal tubule. Ultrastructural segmentation of the proximal tubule was demonstrated in the human fetal nephron in developmental stage IV.     

Clonal development of interconnected germ cells in the rat and its relationship to the segmental and subsegmental organization of spermatogenesis.

Segments and subsegments are the smallest unit of synchrony thus far described within longitudinal sections of seminiferous tubules. It is known that cells in a clone joined by intercellular bridges are at the same phase of development and are also thought to be units of synchrony. This study was designed to determine if it is possible that the synchrony seen in cells joined by intercellular bridges is the same as that cataloged along the long axis of the seminiferous tubule. In the present study, the maximum number of rat spermatids joined by intercellular bridges (a clone) was obtained. It was hypothesized that if the clone size were larger than the smallest known units of synchrony (segments or subsegments) in the long axis of the seminiferous tubule, then intercellular bridges would most likely govern the synchronous development of segments or subsegments (or finer subdivisions thereof). If the clone size is smaller than the number of cells present in a segment or subsegment, then other factors must govern synchrony in the longitudinal aspect of the tubule. In the determination of spermatid clone size, rat testes were injected with cytochalasin D which opens intercellular bridges of a spermatid clone to produce large symplasts. The number of nuclei in the symplasts was determined from serially sectioned tissue, by drawing nuclei with a camera-lucida, and by counting nuclei. After extensive examination of tubules, the number of spermatids found in the suspected five largest clones observed was determined to be 650, 607, 338, 240, and 177.(ABSTRACT TRUNCATED AT 250 WORDS)     

Spontaneous degeneration of germ cells in normal rat testis: assessment of cell types and frequency during the spermatogenic cycle.

The occurrence of degenerating germ cells in the cycle of the seminiferous epithelium was measured in testicular tissues from eight normal adult rats. Testes were perfusion fixed, embedded in epoxy resin and, after sectioning a total of 180 randomly selected blocks at 1 microns, stained sections were examined by light microscopy; all cross-sectioned seminiferous tubules were categorized into one of 14 stages of the spermatogenic cycle. The number of degenerating cells per tubule was recorded in 2103 tubules. Degenerating germ cells were not detected at stages II-VI, and only rarely at stage VII (n = 366 tubules) in which one primary spermatocyte and one step 19 spermatid degenerated. All other stages exhibited a greater incidence of degenerative germ cells, particularly at stage XIV where, on average, the frequency of degenerating cells per round seminiferous tubule was about 40 times greater than at stage VII. The results indicated that, in the normal adult rat testis, the germ cells are least at risk of degeneration as they pass through stage VII.     

Selfish spermatogonial selection: evidence from an immunohistochemical screen in testes of elderly men

The dominant congenital disorders Apert syndrome, achondroplasia and multiple endocrine neoplasia-caused by specific missense mutations in the FGFR2, FGFR3 and RET proteins respectively-represent classical examples of paternal age-effect mutation, a class that arises at particularly high frequencies in the sperm of older men. Previous analyses of DNA from randomly selected cadaveric testes showed that the levels of the corresponding FGFR2, FGFR3 and RET mutations exhibit very uneven spatial distributions, with localised hotspots surrounded by large mutation-negative areas. These studies imply that normal testes are mosaic for clusters of mutant cells: these clusters are predicted to have altered growth and signalling properties leading to their clonal expansion (selfish spermatogonial selection), but DNA extraction eliminates the possibility to study such processes at a tissue level. Using a panel of antibodies optimised for the detection of spermatocytic seminoma, a rare tumour of spermatogonial origin, we demonstrate that putative clonal events are frequent within normal testes of elderly men (mean age: 73.3 yrs) and can be classed into two broad categories. We found numerous small (less than 200 cells) cellular aggregations with distinct immunohistochemical characteristics, localised to a portion of the seminiferous tubule, which are of uncertain significance. However more infrequently we identified additional regions where entire seminiferous tubules had a circumferentially altered immunohistochemical appearance that extended through multiple serial sections that were physically contiguous (up to 1 mm in length), and exhibited enhanced staining for antibodies both to FGFR3 and a marker of downstream signal activation, pAKT. These findings support the concept that populations of spermatogonia in individual seminiferous tubules in the testes of older men are clonal mosaics with regard to their signalling properties and activation, thus fulfilling one of the specific predictions of selfish spermatogonial selection.     

Basal lamina of human seminiferous tubule—Its role in material transport

Summary Electron microscopy of testicular biopsies obtained from two adult males with tunica vaginal hydrocele revealed some protrusions from the basal lamina to the germinal epithelium in the seminiferous tubule. The protrusions were of three types: some between the spermatogonia and Sertoli cells, some directly within the Sertoli cell cytoplasm and others inside the spermatogonia. The protrusions inside the spermatogonia were only 0.5 m deep whereas the other types were from 1–11 m deep. Occasionally some cut off portions of these protrusions were seen inside the ground cytoplasm of the Sertoli cell without an apparent connection with the original stalk. The matrix of the protrusions contained a homogenous component (composed of a fine filamentous element) and granular and membranous components. These components closely resemble the materials found in the basal lamina of the seminiferous tubule. It has been suggested that under mild pathological conditions, i.e., hydrocele, the junctions between the seminiferous tubule epithelium and the basal lamina become somewhat more flexible. As a result, the protrusions become longer and a passage might be formed to allow the flow of raw materials in or out of the seminiferous tubule.This work was supported by USPHS Research Grant HD-03266 and GRS 94802     

Effects of prenatal exposure to a 900 MHz electromagnetic field on 60-day-old rat testis and epididymal sperm quality

We investigated the effects of exposure in utero to a 900 megahertz (MHz) electromagnetic field (EMF) on 60-day-old rat testis and epididymis. Pregnant rats were divided into control (CG; no treatment) and EMF (EMFG) groups. The EMFG was exposed to 900 MHz EMF for 1 h each day during days 13 ? 21 of pregnancy. Newborn rats were either newborn CG (NCG) or newborn EMF groups (NEMFG). On postnatal day 60, a testis and epididymis were removed from each animal. Epididymal semen quality, and lipid and DNA oxidation levels, apoptotic index and histopathological damage to the testis were compared. We found a higher apoptotic index, greater DNA oxidation levels and lower sperm motility and vitality in the NEMFG compared to controls. Immature germ cells in the seminiferous tubule lumen, and altered seminiferous tubule epithelium and seminiferous tubule structure also were observed in hematoxylin and eosin stained sections of NEMFG testis. Nuclear changes that indicated apoptosis were identified in TUNEL stained sections and large numbers of apoptotic cells were observed in most of the seminiferous tubule epithelium in the NEMFG. Sixty-day-old rat testes exposed to 900 MHz EMF exhibited altered sperm quality and biochemical characteristics.     

Dolichol and N-linked oligosaccharide synthesis in the rat testis: interaction between Sertoli and spermatogenic cells, evidence for paracrine effects.

The relative contribution of the Sertoli cell and the pachytene spermatocyte to dolichol and N-linked oligosaccharide biosynthesis within the seminiferous tubule was investigated. Evidence is presented to show that the interaction between these two cell types affects dolichol and N-linked oligosaccharide biosynthesis. Analysis of the dolichol content of Sertoli cultures confirms earlier data suggesting that the Sertoli cell constitutes the major pool of dolichols within the seminiferous tubule. [14C]Acetate incorporation studies suggest that the Sertoli cell in culture synthesizes dolichol much more rapidly than does the isolated pachytene spermatocyte. This information, in addition to previous data in the literature, infers an interactive effect whereby the presence of the spermatogenic cell in the tubule stimulates dolichol synthesis in the Sertoli cell. The absence of normal Sertoli-spermatocyte interactions in in vitro incubations may also limit dolichol synthesis in the pachytene spermatocyte. The distribution of dolichol kinase between the Sertoli and the pachytene spermatocyte was also examined. The concentration of this enzyme in the Sertoli cell suggests the presence of an active salvage pathway within that cell. The correlation between the appearance of the pachytene spermatocyte and the previously described peak of dolichol kinase activity in the seminiferous tubules of the prepubertal animal implies cell-cell interactions. Radiolabelling studies of N-linked oligosaccharides were conducted using [3H]mannose and concanavalin A affinity chromatography to identify multiantennary, biantennary, and high-mannose oligosaccharide pools. An in vitro bicameral coculture system was used to demonstrate that pachytene spermatocytes stimulate incorporation of [3H]mannose into Sertoli cell oligosaccharides.(ABSTRACT TRUNCATED AT 250 WORDS)     

The shape and distribution of lysosomes and endocytosis in the ciliary epithelial cells of rats

Stages of the spermatogenic cycle in human seminiferous tubules were evaluated in men with varied efficiencies of spermatogenesis to determine if the architectural arrangement of stages or the atypical cell types contributed to variation in sperm production rates. Testes were selected from men with low, intermediate, and high daily sperm production per g parenchyma (DSP/g). Round tubular cross sections were photographed by bright-field microscopy. Stages were identified for each cross section by two observers and the number of stages represented in each cross section was recorded. Number of stages per cross section in men with low efficiency of spermatogenesis were significantly (P<0.05) fewer than men with intermediate and high efficiency of spermatogenesis. Further, the percentage of stages with atypical cell types in men with high DSP/g was significantly (P<0.05) higher than men with low DSP/g. There was a significant relationship (P<0.01) between the percentages of stages with atypical cell types per stage and number of stages per cross section. The atypical cell types appear to result from high density of stages per cross section in men with high DSP/g. There was no significant difference observed between groups for tubular volume, diameter, length, volume density, and volume density of seminiferous epithelium. However, a significant (P<0.05) positive correlation between percent seminiferous epithelium per testis with DSP/g or with the number of stages per cross section was found. These findings reveal that the architectural makeup of stages within seminiferous tubules and atypical cell types within stages varies with the level of efficiency of spermatogenesis, and this variation may reflect differences in yield of early spermatogonial divisions that are responsible for generating the different stages.     

Prognostic role of natural killer cells in pediatric mixed cellularity and nodular sclerosing Hodgkin's disease

OBJECTIVE: To study natural killer cells' spontaneous cytotoxic capacity against tumor cells and their prognostic significance in classical Hodgkin's disease. STUDY DESIGN: Thirty-eight pediatric mixed cellularity and nodular sclerosing Hodgkin's disease patients were included in the study. Immunohistochemical staining was performed for natural killer cells in the background using the monoclonal antibody CD57 in serial sections of B5-formalin-fixed, paraffin-embedded tissue blocks. CD57-positive cells were counted with an immersion objective among 5,000 cells on representative areas of the tumors. The degree of natural killer cells was classified as low (<150 cells) or high (> or = 150 cells). Multivariate regression analysis was performed to determine the differences between patients with and without relapse. RESULT: The mean of CD57-positive cell numberfor all the cases was 173.42 +/- 117.34 (range, 20-500). CD57-positive cells were high in 21 cases and low in 17. The mean of CD57-positive cell numbers was 191.85 +/- 115.33 in the disease-free group and 84.44 +/- 57.90 in the relapsing group. Log rank analysis showed statistical significance between event-free survival and number of CD57-positive cells (P = .0207). CONCLUSION: In multivariate analysis, CD57 expression proved to be a prognostic factor independent from otherfactors. As a result, CD57 expression by background natural killer cells may be used as a prognostic parameter in mixed cellularity and nodular sclerosing Hodgkin's disease.