Alternate pathogenesis of systemic neoplasia in the bivalve mollusc Mytilus

JOURNALOF

INVERTEBRATEPATHOLOGY

Alternate

58,231-243

(1991)

Pathogenesis of Systemic Neoplasia Bivalve Mollusc Mytiks

J. D. MOORE, R. A. ELSTON, A. S. DRUM,AND Battelle

Marine

Sciences

Laboratory,

439 West Sequim

Bay Road,

M.T.

in the

WILKINSON

Sequim,

Washington

98382

Received October 1I, 1990; accepted December 13, 1990 The proliferative disease systemic neoplasia, also termed hemic neoplasia or disseminated sarcoma, was studied in four Puget Sound, Washington populations of the bay mussel (Mytilus sp.). Using flow cytometric measurement of DAPI-stained cells withdrawn from the hemolymph, DNA content frequency histograms were generated for 73 individuals affected by the disease. The cells manifesting systemic neoplasia were found to exist as either of two separate types, characterized by G,,G, phase nuclear DNA contents of either approximately 4.9~ haploid (pentaploid form) or approximately 3.8~ haploid (tetraploid form). The two disease forms were found to coexist in all four mussel populations sampled, with overall relative prevalences of 66% pentaploid form, 2% tetraploid form, and 5% exhibiting both disease forms simultaneously. These findings represent the first unequivocal demonstration of multiple cell types in a bivalve neoplasia. The two forms appear to represent separate pathogenetic processes rather than sequential stages of a single pathogenesis. Two cell cycling parameters associated with proliferative activity were employed to compare the alternate forms: (i) the percentage of cells assigned to the DNA Synthesis (S) phase of the neoplastic cell cycle, and (ii) the proportion of neoplastic cell mitotic figures in hemocytological preparations. Mean values for both parameters were significantly higher for mussels with the tetraploid form of the disease, suggesting a higher rate of proliferation relative to the pentaploid form. Qualitatively, cells of the tetraploid form contained slightly lower nuclear and cytoplasmic volumes compared to those of the pentaploid form. An observed wide variation in neoplastic cell nuclear size within either disease form may reflect the distribution of cells in the G,G,, S, and G,M phases of the cell cycle. Potential etiologic relationships between the two forms are discussed. 0 1991 Academic Press, Inc. KEY WORDS: neoplasia; cancer; Mytilus; bivalve; mollusc; flow cytometry; polyploid; cell cycle; proliferation rate.

INTRODUCTION

Mytilus (Elston et al., 1988a) and M. arenuriu (Smolowitz et al., 1989).

Systemic proliferation of neoplastic cells in bivalve molluscs has been reported for at least 15 species worldwide (Peters, 1988). The systemic neoplasia of Puget Sound, Washington Mytilus bay mussels has been shown to be progressive, fatal (Elston et al., 1988a), and transmissible through a cell homogenate (Elston et al., 1988b). These observations suggest that neoplastic transformation is initiated by an infectious etiologic agent, and a retroviral etiology has been claimed for a similar condition in the soft-shell clam Mya arenuriu (Oprandy et al., 1981; Oprandy and Chang, 1983). The tissue from which the neoplastic cells are derived remains unknown in all bivalves studied, although indirect evidence points to a hemic origin for the condition in

Morphologically, the systemic neoplasia of Puget Sound Myths is manifested by massive proliferation of atypical cells with scant amounts of cytoplasm, large nuclei with prominent nucleoli, and high proportions of mitotic figures. Flow cytometric measurement of DNA content has recently been employed to follow progression of the disease in individuals held in a laboratory setting (Elston et al., 1990). The neoplastic cells in the majority of mussels were characterized by a unique cell cycle with a G,,G, phase DNA content of approximately 5x haploid (5n; pentaploid) and passage through a DNA Synthesis (S) phase to a G,/Mitosis (G,M) phase of approximately lOn, presumably followed by division and repetition of the cell cycle. Surprisingly, 2 231 0022-2011/91 $1.50 Copyright All rights

8 1991 by Academic Press, Inc. of reproduction in any form reserved.

232

MOORE

of 52 individuals showed an apparently similar proliferative pattern with a G,G, phase DNA content of approximately 4x haploid (4n; tetraploid). This finding suggested the existence of a separate form of neoplasia, though previous morphological examination had not detected two cytologic types. The potential existence of two pathogenetic forms of systemic neoplasia is of importance with respect to the topics of tissue of origin, mechanism of neoplastic transformation, and etiology. In this paper we present unequivocal documentation of the cooccurrence of alternate forms of systemic neoplasia within Mytilus populations and, rarely, within individuals. The two forms appear to represent distinct pathogenetic processes rather than sequential stages of a single pathogenesis. In addition to characterizing the alternate cell cycles, cellular morphologies are compared and potential etiologic relationships between the two forms are discussed. MATERIALS AND METHODS Collection of Population Samples Mytilus sp. bay mussels, shell length 3185 mm, were collected from four populations in the Puget Sound area of Washington, between March, 1989 and January, 1990 (Poulsbo Marina, Penn Cove on Whidbey Island, Sequim Bay, and Port Angeles Harbor). The historical classification of Puget Sound bay mussels as Mytilus edulis (based on morphometry) has been questioned by a recent isozyme study which concluded a species assignment of Myths trossulus to these populations (McDonald and Koehn, 1988). The Poulsbo Marina mussels were collected from an area of less than 1 m* on a single pier piling, and were considered representative of mussels in close proximity. The Sequim Bay and Port Angeles Harbor mussels were each collected from several adjacent pier pilings, while those from Penn Cove were collected from various raft lines of a commercial mussel farm which utilizes set from natural spawnings.

ET

AL.

Rapid Screening for Systemic Neoplasia The entire collection of Poulsbo Marina Myths (n = 60) was processed in order to obtain asymptomatic in addition to neoplasia-positive individuals. Penn Cove, Sequim Bay, and Port Angeles collections of several hundred mussels each were screened for systemic neoplasia by macroscopically examining withdrawn hemoiymph. A 22-gauge needle was inserted between the valves and into the posterior adductor muscle, followed by withdrawal of 0.2-0.5 ml of hemolymph into a syringe containing 0.5 ml of cold hemocyte buffer (0.01 M tris, 6 mM EDTA, 860 mOsm NaCl, pH 8.4; after Wittke and Renwrantz, 1984). Individuals without systemic neoplasia or with a low level of either disease form had white hemolymph which mixed rapidly with the buffer. The hemolymph of mussels in later stages of either form of systemic neoplasia was copious, often yellowish in color, and sank to the bottom of the collection syringe. Mussels exhibiting this characteristic hemolymph were processed for further analysis as described below. Hemocytology Hemolymph samples obtained as described above were prepared for microscopic examination (hemocytology) and for measurement of the distribution of nuclear DNA contents (flow cytometry). For hemocytology, 0.1 ml of hemolymph-buffer solution was placed on a poly-L-lysine coated microscope slide, incubated in a moist chamber for 20 min at room temperature, and stained by a Fuelgen-picromethyl blue procedure (after Farley, 1969). Flow Cytometry Sample preparation and processing for flow cytometry followed the procedures of Elston et al. (1990). Briefly, OS-O.7 ml of hemolymph-buffer solution was centrifuged (3OOg, 2 min, 4°C) within ten min of hemolymph withdrawal and the resulting cell pellet was resuspended in 1 ml of a DAPI

ALTERNATE

PATHOGENESIS

OF Mytilus

NEOPLASIA

233

Cell Cycling Parameters The pentaploid and tetraploid disease forms were compared using two parameters associated with cellular proliferative activity: (i) the percentage of neoplastic cells assigned to the S phase compartment of the neoplastic cell cycle (% S phase), and (ii) the estimated proportion of neoplastic cell mitotic figures in hemocytological preparations. Samples for which the proportion of neoplastic cells was less than 50% of the total cells present were excluded from these comparisons to minimize variation in disease severity. A two-sample t test with (Y = 0.05 (Zar, 1984) was employed for each comparison. Proportions of neoplastic cells with miCell Cycle Analysis totic figures were estimated by scanning Each DAPI-stained hemolymph sample well-monolayered hemocytological prepawas processed by flow cytometry twice; rations and noting the number of mitotic first as a pure sample to obtain an unaltered figures (prophase to telophase) present. No DNA content histogram (Figs. l-3), and distinction was made between normal and subsequently with the addition of a mussel abnormal mitotic figures. Twenty-five ransperm standard for consistent identification dom fields of view at 500x magnification of the ploidy level of histogram peaks (El- were examined for each sample hemocytoston et al., 1990). logical preparation. The total number of Cell cycling parameters were calculated neoplastic cells per field of view varied befrom DNA content frequency histograms tween views; to account for this, the prowith interactive “MultiCycle” software portion of the field of view covered by neo(Phoenix Flow Systems, San Diego, Caliplastic cells (% cover) was estimated and fornia), which assigns cells in a DNA con- transformed into the number of neoplastic tent histogram to the GoGl, S, or G,M cells examined per view, using estimates of phases of the cell cycle or, when present, the number of cells present in a full monomultiple cell cycles. For each sample the layer (100% cover; 2270 and 3343 cells for diploid histogram peak was defined as hav- the pentaploid and tetraploid forms, respecing a relative DNA content value of 2.0X tively). The sum of neoplastic cell mitotic haploid. Histogram peaks representing figures in 25 fields of view was divided by GoGl phase neoplastic cells were converted the sum of all neoplastic cells in the 25 from channel numbers to values on a scale fields viewed, and this value was treated as relative to haploid DNA content by divida single sample data point in statistical analing the modal (peak) channel number by ysis. one-half that of diploid. Those histograms Histology which had a diploid or neoplastic G,G, After obtaining hemolymph samples, peak coefficient of variation greater than 5 were considered imprecise and were ex- whole animals were shucked and a cross section adjacent to the foot was fixed in cluded from cell cycle analyses (after MerDavidson’s solution. After embedding in kel et al., 1987), but were used for estimaand eosin-stained tions of relative prevalence of the two dis- paraffin, hematoxylin6-km sections were produced. ease forms.

(4’,6-diamidino-2-phenylindole dihydrochloride) fluorochrome solution including 0.12% w/v Triton X-100 to remove most of the cytoplasm from each cell. DAPI binds specifically to DNA and the fluorescence intensity of a DAPI-stained nucleus is directly proportional to its DNA content (Shapiro, 1988). Samples were held at - 70°C until processing with a Phwye ICP22 flow cytometer at a wavelength of 365 nM. The fluorescence intensities of approximately 15,000 DAPI-stained nuclei per mussel hemolymph sample were measured and placed into a framework of 256 “channel numbers” of relative DNA content, creating a DNA content frequency histogram.

234

MOORE ET AL.

RESULTS Pentaploid and Tetraploid Systemic Neoplasia

TABLE

Forms of

The systemic neoplasia of Puget Sound Mytilus populations exists in two forms with distinct pathogeneses. The alternate forms are characterized by proliferation of either of two neoplastic cell types, with mean G,,Gi phase DNA contents of approximately 4.9 and 3.8~ haploid (Table 1). Because these values are near 5x haploid DNA content and 4~ haploid DNA content, the two disease forms have been termed pentaploid and tetraploid, respectively. These ‘DNA ploidy’ levels refer solely to DNA content; inferences regarding chromosome number or structure cannot be made without applying specific karyological techniques. Figure 1 demonstrates the ploidy levels and proliferation cycles of normal diploid hemocytes and the two forms of neoplasia. In mussels diagnosed as asymptomatic for neoplasia by histology and hemocytology, approximately 92% of the circulating cells were diploid (2n DNA content; Fig. IA), constituting the G,G, compartment of the normal cell cycle. The remaining cells were divided between the S and G,M (4n DNA content) phases of the diploid cell cycle. Circulating cells with DNA content greater than 4n were not seen in asymptomatic mussels. In the pentaploid form of systemic neoplasia, cells of approximately 5n DNA content arise and proliferate through a 5 to 10n DNA content cell cycle (Fig. lB), as previously described by Elston et al. (1990). The tetraploid form of the disease is characterized by an analogous cell cycle for which the G,G, phase DNA content is approximately 4n (Fig. 1C). The pentaploid and tetraploid forms of systemic neoplasia were both seen within all four population samples (Table 2), including that from Poulsbo Marina which was made up of mussels collected within an area of less than 1 m2. Cumulative, relative prevalences over all four population samples were 66% pentaploid form, 29% tetra-

1

DNA CONTENTOFG~G~

PHASECELLSOFNORMAL AND NEOPLASTIC CELL CYCLES

Cell cycle

N

G,G, DNA content”.’

Range

Diploid’ Pentaploid Tetraploid

8 33 17

2 * 0.1 4.9 f 0.1 3.8 * 0.1

1.9-2.1 4.6-5.2 3.M

LIMean values expressed relative to haploid DNA content, 2 SD. b Samples for which the coefficient of variation of the normal or neoplastic GaG, phase peak was greater than 5% were excluded. ’ Diploid cell DNA content defined as 2x haploid. Data from asymptomatic individuals.

ploid form and 5% exhibited both pathogenetic forms simultaneously. Mussels with either disease form exhibited stages of extremely advanced disease progression in which neoplastic cells constituted up to 98.8% (pentaploid form) or 98.6% (tetraploid form) of the circulating cells present. Pathologic characteristics of each form were similar to those reported for the disease in general (Elston et al., 1988a) and for other bivalve systemic neoplasias. These features include proliferation of neoplastic cells throughout the open circulatory system, displacement of normal tissue, and tissue necrosis in extremely advanced stages. S Phase Percentage The pentaploid and tetraploid forms of systemic neoplasia were found to have significantly different (P < 0.001) mean percentages of cells in the S phase of their respective cell cycles (Fig. 2, Table 3). The mean S phase percentage for the tetraploid form was over three times higher than that for the pentaploid form, suggesting a higher rate of proliferation in the tetraploid form. Figure 3 presents a cell cycling model for a mussel in the rare condition of exhibiting both disease forms simultaneously. The two neoplastic cell cycles appear similar to those seen in mussels with single disease forms, including the percentage of cells in the S phase of each cell cycle.

ALTERNATE

PATHOGENESIS

235

OF Mytilus NEOPLASIA

Relative DNA Content (x haploid)

L.”

7.”

“.”

810

Id.0

Relative DNA Content (x haploid) 2800 k D

2400 -

C

Relative DNA Content (x haploid) FIG. 1. Patterns of DNA content in normal and neoplastic Myth circulating cells. DNA ploidy of histogram peaks calculated after identification of the diploid peak using a Myths sperm standard (not shown). Note that ordinate and abscissa scales are not constant. (A) DNA content frequency histogram of a mussel asymptomatic for systemic neoplasia by morphological criteria. Large peak at 2n DNA content represents diploid cycle G,G, cells. Small diploid cycle G,M peak at 4n DNA content is also apparent. (B) Typical DNA content frequency histogram of a mussel with advanced pentaploid form of systemic neoplasia. Normal diploid cycle cells are overwhelmed in numbers by the neoplastic cells of 5n (neoplastic cycle G,G,) to 10n (neoplastic cycle G,M) DNA content. (C) Typical DNA content frequency histogram of a mussel with advanced tetraploid form of systemic neoplasia. GaG, phase cells of the neoplastic cell cycle are near-tetraploid in DNA content and cycle to a G,M phase of approximately 8n DNA content.

Mitotic

Figures

Neoplastic cell mitotic figures were numerous in mussels with either form of systemic neoplasia. Comparing the alternate

forms, the mean proportion of mitotic figures was significantly different (P < O.OOl), that of the tetraploid form being nearly five times higher than that of the pentaploid form (Table 4).

236

MOORE TABLE

2

RELATIVEPREVALENCEOFPENTAPLOIDAND TETRAPLOID FORMS OF SYSTEMIC NEOPLASIA IN FOUR Myths POPULATIONS Relative prevalence Population Poulsbo (n = 3) Sequim Bay (n = 17) Penn Cove (n = 7) Port Angeles (n = 46) Cumulative (n = 73, 100%)

Pentaploid form

Tetraploid form

Both”

2

1

0

15

2

0

3

4

0

28

14

4

48 (66%)

21 (2%)

4 (5%)

u Simultaneous presence of pentaploid ploid form cells in individual mussels.

and tetra-

Qualitative Comparison of Neoplastic Cell Morphology

Subtle differences in cell morphology between the two pathogenetic forms were observed in hemocytological and histological preparations (Figs. 4,5). Cells of the pentaploid form typically had extremely large nuclei of variable size, prominent single nucleoli, and scanty cytoplasm (Figs. 4A,5A). Spreading of cytoplasm on the poly-Llysine coated slides of hemocytological preparations was minimal and pseudopodia formation was either absent or consisted of short, unidirectional cytoplasmic extensions. Cells of the tetraploid form exhibited similar behavior and appearance, yet had a higher prevalence of double nucleoli and generally smaller nuclear and cellular volumes (Figs. 4B,5B). A wide range in neoplastic cell nuclear size was observed in individuals with either disease form. Considering the proportions of cells in the S + G,M phases of the neoplastic cell cycles, this pleomorphy may reflect the active cycling of neoplastic cell populations (see Discussion). DISCUSSION

Systemic neoplasia in Puget Sound exists in two pathogenetic forms distinguished by cell cycles with clearly

Mytilus

ET AL.

separate DNA ploidy levels. The DNA content of G,G, phase neoplastic cells is either approximately 5x haploid or 4x haploid; rarely, both types of cells occur simultaneously in a single mussel. It must be emphasized that euploid DNA content does not imply karyological euploidy, as minor numerical alterations and any structural alterations from the euploid karyotype are impossible to detect by the method of DNA ploidy measurement used in this study. The existence of alternate forms of systemic neoplasia in Puget Sound Mytilus went unrecognized prior to the application of flow cytometry. This was due to similar cytomorphological and pathological characteristics between the pentaploid and tetraploid forms. These shared characteristics (high nucleus to cytoplasm volume ratio, pleomorphy, polyploidy, prominent nucleoli, frequent mitotic figures, displacement and destruction of normal tissues) are typical features of rapidly proliferating neoplastic cells (Pastan and Willingham, 1978; Robbins and Kumar, 1987). The similarities are therefore not unexpected, and support the premise that both forms represent truly neoplastic cell populations. Alternate Pathogenesis

The pentaploid and tetraploid forms of systemic neoplasia are apparently independent once established in individual mussels. Previous laboratory studies followed progression of the pentaploid form in individual mussels from nonclinical stages to death (Elston et al., 1990). Massive proliferation of pentaploid form cells was accompanied by no change in the normal, low percentage of tetraploid cells (which represent the G,M phase of the diploid hemocyte cell cycle). In the present study, mussels in extremely advanced stages of either form were observed, with moribund pathological indications. The pentaploid and tetraploid forms therefore appear to represent distinct pathogenetic processes rather than sequential stages of a single pathogenesis. Hyperdiploid cell lines with GoG, phase DNA contents other than approximately 4n

ALTERNATE

g

160

z’

120

PATHOGENESIS

OF Mytilus

NEOPLASIA

237

5 aa 2 3 2

80 40

Relative DNA Content (x haploid) 450 400 350 300 250 200 150 100 50 0

2.0

4.0

5.0

8.0

10.0

Relative DNA Content (x haploid) 280 k

240

f z' =

200

3 Q) 2 ;ii d

160 120 80 40 0

4 2.0

4.0

5.0

8.0

Relative DNA Content (x haploid) FIG. 2. Cell cycling models for the DNA content histograms shown in Fig. 1. Ordinate scales have been expanded 10x. Shown are the assignment of cells of various DNA contents to f&G,, S, and G,M phases of cell cycles, S phases shaded. (A) Asymptomatic mussel with low proportions of cells in S and G,M phases of the diploid cell cycle. (B) Mussel with advanced pentaploid form systemic neoplasia. 16% of the neoplastic cells were assigned to S phase. (C) Mussel with advanced tetraploid form systemic neoplasia. 30.6% of the cells of the neoplastic cycle were assigned to S phase, suggesting a higher rate of proliferation relative to the pentaploid form.

and approximately 5n have not been associated with the Mytifus neoplasia. This consistency suggests that neoplastic transformations giving rise to each form are associated with specific chromosomal aberrations, as reported for many mammalian leukemias and lymphomas (Robbins and Kumar, 1987). Various leukemias have

been shown to have a retroviral etiology in which the virus integrates into the host genome at numerous sites, one to several of which can result in neoplastic transformation as, for example, shown for the bovine leukemia virus (Burny et al., 1988). It is conceivable that the pentaploid and tetraploid forms of Myths systemic neoplasia

238

MOORE ET AL. TABLE

COMPARISON CELLS

3

TABLE

OF THE PERCENTAGE OF NEOPLASTIC IN S PHASE OF PENTAPLOID AND TETRAPLOID CELL CYCLES

% Cells in S phase of neoplastic cell cycle” Pentaploid form Mean 2 SD Range; n t test

Mitotic figures per 1000 neoplastic cells”

Tetraploid form

11.2 f 6.1 38.1 + 9.6 2.1-23.7; 31 15.g52.2; 16 P < 0.001

0 Samples for which the coefficient of variation of the normal or neoplastic G,G, phase peak was greater than 5%, or for which the proportion of neoplastic cells was less than 50%, were excluded.

could arise in such a manner. Under this model, infectious particles released from mussels with either disease form would be identical, and could result in either disease form in recipient mussels. Mussels with simultaneous cooccurrence of the two forms may represent rare occasions when independent transformations occur at approximately the same time. Such cases of “multiploidy” are well documented in vertebrate neoplasias (Merkel et al., 1987). A second alternative is that a single etiologic agent is capable of transforming two wholly separate tissue types. It is interesting to note that a murine leukemia virus has been shown to be capable of transforming reproductive tissue in addition to hemopoi-

4

COMPARISON OF THE PROPORTION OF NEOPLASTIC CELL MITOTIC FIGURES IN HEMOCYTOLOGICAL PREPARATIONS FROM MUSSELS WITH THE PENTAPLOID AND TETRAPLOID FORMS OF SYSTEMIC NEOPLASIA

Pentaploid form Mean -+ SD Range; n t test

Tetraploid form

0.64 k 0.65 O-2.38;

2.95 * 0.47 2.22-3.48; 5

12

P < 0.001

0 Data from samples with well-monolayered hemocytological preparations and for which the proportion of neoplastic cells was greater than 50%.

etic tissue (Pierpaoli and Meshorer, 1982), and that presumptive hemopoietic and gonadal neoplasms have been reported to cooccur within specific bivalve populations, notably for the soft-shell clam M. arenaria, (Brown et al., 1977; Yevich and Barszcz, 1977) and British Columbia M. edufis (Cosson-Mannevy et al., 1984). Finally, the alternate disease forms could derive from two completely independent (although perhaps closely related) etiologic agents. Cell Cycling of Alternate Forms

The proportion of neoplastic cell mitotic figures and the percentage of cells in S

320

& 280 2 3 3

240

.-9 iii 2

120

200 160

60 40 Relative DNA Content (x haploid)

FIG. 3. Cell cycling model showing the rare occurrence of a mussel with simultaneous presence of the pentaploid and tetraploid forms of systemic neoplasia. Peak frequencies near 4n and 5n DNA content represent G,G, phase cells of tetraploid and pentaploid cycles, respectively.

FIG. neoplas nucleus volume

4.

Hemocytological

Tetraploid (2 u-rows). ra tios than

ia.

systemi IC 1reoplasia cell cyc :le. Bar =

of the

pentaploid

form cells generally have higher proportion of mitotic pentaploid form cells. Broad

comparison

smaller figures ranges

may be explained IO pm.

in part

by the 239

(A)

and

tetraploid

(B)

forms

of systc :mic

nuclei, higher prevalence of two nucleoli per asm (arrowhead). and higher nucleus:cytopl in nuclear size in the cells of both form IS of presence of cells in the S and G,M phases 01 F the

MOORE ET AL.

FIG. 5. Histological comparison of the pentaploid (A) and tetraploid (B) forms of systemic neoplasia. Note the relative size of gill epithelial nuclei in (B). Bar = 10 pm.

phase of the neoplastic cell cycle were estimated for the alternate disease forms. Values for each parameter were much higher than for normal circulating cells,

providing strong evidence that the neoplastic cells of each disease form replicate in circulation. Mix (1973, working with the atypical cells of the “presumptive neoplas-

ALTERNATE

PATHOGENESIS

tic disease” of the oyster Ostreola con(formerly Ostrea lurida) from Yaquina Bay, Oregon, similarly found that incorporation of tritiated thymidine was higher in the atypical cells than in any normal tissue and concluded that the cells undergo rapid proliferation. The S phase percentage is considered an indicator of cycling activity on the premise that a cell population proliferates at a rate dependent on the fraction of cycling cells, while the cycle duration time remains constant (Pardee, 1989). Numerous studies have reported a good correlation between the S phase percentage measured by flow cytometry and that measured by incorporation of tritiated thymidine (e.g., McDivitt et al., 1985; Braylan et al., 1980). However, the cycle duration time may not always be uniform (Gray et al., 1979; Rabinovitch et al., 1988) and quiescent S and G,M phases of transformed cells have been identified (Shapiro, 1988). Though rarely observed, such findings limit the confidence with which the S phase percentage can be considered a quantitative indicator of cycling activity. Use of the proportion of mitotic figures as a measure of cycling activity is limited by the analogous potential for mitotic arrest. Merkel and colleagues (1987) reviewed the use of cell cycle analysis as a prognostic tool for human malignancies. In studies involving patients with non-Hodgkins lymphoma, leukemias, and solid tumors, a higher proportion of cells assigned to S phase of the neoplastic cell cycle correlated with “unfavorable” prognosis by standard techniques, and with shorter survival time. Virtually all studies reported a wide variation of S phase percentages even within subcategories of specific tumors, in agreement with our findings of wide variation of S phase percentages within the pentaploid and tetraploid forms of systemic neoplasia (Table 3). In the present study, comparisons of the S phase percentage and the proportion of mitotic figures each suggest that the cycling activity of the tetraploid disease form is higher than that of the pentaploid chaphila

OF

Mytilus

NEOPLASIA

241

form. The tetraploid form may therefore exhibit faster disease progression and a higher mortality rate. Under these conditions, point-in-time prevalence data may lead to underestimation of the relative impact of the tetraploid form on mussel population mortalities. Size Variation

and the Cell Cycle

A wide range in neoplastic cell size was observed in individuals with either form of systemic neoplasia. Such pleomorphy may reflect the distribution of neoplastic cells in the G,G,, S, and G,M phases of the cell cycle. Mix (1975) described two intraindividual types of atypical cells within 0. conchaphila, noting that a continuum of morphologies between the two was also observed. The more abundant cells were smaller and had smaller nuclei, as would be expected of cells in the G,G, phase of the neoplastic cell cycle. The less abundant, larger cell type may have been the G,M phase of the same cell population. Lowe and Moore (1978) provide evidence for a correlation between DNA content and size in the neoplastic cells of United Kingdom M. edulis. Measuring the size and DNA content of individual nuclei, they reported two intraindividual neoplastic cell types, again with a continuum between them. The more abundant type was smaller, with irregular nuclei, multiple nucleoli, and DNA content 3.8-4.3x diploid; a second, less frequent type was larger, with DNA content estimated at 6.8-7.7x diploid. The authors note that the second type could represent a “premitotic step” of the first. Correlation of cell size and DNA content has also been reported in a study on normal proliferating human bone marrow cells (Barlogie et al., 1980). In asynchronous populations of proliferating cells, distribution by DNA content is extremely non-Gaussian (Fig. 1). If distribution by size is a function of DNA content, then arithmetic means and variances of neoplastic cell size, as widely reported in the literature on systemic neoplasia, are in-

242

MOORE

appropriate because of the underlying sumption of a Gaussian distribution.

as-

Detection of Systemic Neoplasia and Discrimination of the Pentaploid and Tetraploid Forms

Flow cytometry provided a rapid method of measuring the DNA content of thousands of hemolymph cells per individual mussel. The technique allowed for objective, unequivocal detection and discrimination of the two disease forms, as long as the neoplastic cell population was large enough for detection on a DNA content histogram. Our unpublished observations indicate that unequivocal detection of the pentaploid form of systemic neoplasia is possible when the proportion of cells in the G,G, phase peak (5n DNA content) reaches approximately 2-3% of the total cells present. Identification of the tetraploid form is dependent on detection of an abnormally high proportion of tetraploid DNA content cells and the appearance of a 4n-8n DNA content cell cycle. In early stages of disease progression, systemic neoplasia can be detected at lower levels by either histology or hemocytology than by flow cytometry; neoplastic cells can be identified at much lower frequencies by microscopic examination than by flow cytometry (unpubl. observations). However, discrimination of disease form as pentaploid or tetraploid utilizing microscopic preparations is difficult, due to the aforementioned similar appearance and highly variant, overlapping size distributions. In summary, two distinct forms of systemic neoplasia exist within Mytilus populations, identifiable by their clearly separate cell cycles. They represent distinct pathogenetic processes and appear to differ in rate of proliferation. Further studies are needed to determine molecular bases of transformation and the etiologic relationship between the two forms. ACKNOWLEDGMENTS This work was supported by the National Cancer

ET AL.

Institute and the US Army Medical Research and Development Command under Grant No. SRC (8), 5 ROl CA 44269-02, and the US Department of Agriculture under Grant No. 88-34123-3685.

REFERENCES BARLOGIE, B., LATREILLE, J., FREIREICH, E. J., Fu, C., MELLARD, D., MEISTRICH, M., AND ANDREEFF, M. 1980. Characterization of hematologic malignancies by flow cytometry. Blood Cells, 6, 719-744. BRAYLAN, R. C., DIAMOND, L. W., POWELL, M. L., AND HARTY-GOLDER, B. 1980. Percentage of cells in the S phase of the cell cycle in human lymphoma determined by flow cytometry. Cytometry, 1, 171174. BROWN, R. S., WOLKE, R. E., SAILA, S. B., AND BROWN, C. W. 1977. Prevalence of neoplasia in 10 New England populations of the soft-shell clam (Mya 534.

arenaria).

Ann.

N.Y.

Acad.

Ski.,

298, 522-

BURNY, A., CLEUTER. Y., KETTMANN, R., MAMMERICKX, M., MARBAIX, G., PORTETELLE, D., VAN DEN BROEKE, A., WILLEMS, L., AND THOMAS, R. 1988. Bovine leukaemia: Facts and hypotheses derived from the study of an infectious cancer. Vet. Microbiol., 17, 197-218. COSSON-MANNEVY. M. A., WONG, C. S., AND CRETNEY, W. J. 1984. Putative neoplastic disorders in mussels (Myths edulis) from Southern Vancouver Island waters, British Columbia. J. Invertebr. Parhol., 44, 151-160. ELSTON. R. A., KENT, M. L., AND DRUM, A. S. 1988a. Progression, lethality and remission of hemic neoplasia in the bay mussel Mytilus edulis. Dis. Aquat. Org., 4, 135-142. ELSTON, R. A., KENT, M. L., AND DRUM, A. S. 1988b. Transmission of hemic neoplasia in the bay mussel, Myths edulis, using whole cells and cell homogenate. Dev. Camp. Immunol., 12, 719-727. ELSTON, R. A.. DRUM, A. S., AND ALLEN, S. K. JR. 1990. Progressive development of circulating polyploid cells in Myths with hemic neoplasia. Dis. Aquat. Org., 8, 51-59. FARLEY. C. A. 1969. Probable neoplastic disease of the hematopoietic system in oysters, Crassostrea virginica and Crassostrea gigas. Natl. Cancer Inst. Monogr.,

31, 541-555.

GRAY, J. W., DEAN, P. N., AND MENDELSOHN. M. L. 1979. Quantitative cell cycle analysis. In “Flow Cytometry and Sorting” (M. R. Melamed, P. F. Mullaney, and M. L. Mendelsohn, Eds.), pp. 383-407. Wiley. New York. LOWE, D. M.. AND MOORE, M. N. 1978. Cytology and quantitative cytochemistry of a proliferative atypical hemocytic condition in Myths edulis (Bivalvia, Mollusca). J. Natl. Cancer Inst., 60, 14551459.

MCDIVITT, R. W., STONE, K. R., CRAIG, R. B., AND MEYER, J. S. 1985. A comparison of human breast

ALTERNATE

PATHOGENESIS

cancer cell kinetics measured by flow cytometry and thymidine labeling. Lab. Invest., 52, 287-291. MCDONALD, J. H., AND KOEHN, R. K. 1988. The mussels Mytilus gailoprovincialis and M. trossulus on the pacific coast of North America. Mar. Biol., 99, 111-118. MERKEL, D. E., DRESSLER, L. G., AND MCGUIRE, W. L. 1987. Flow cytometry, cellular DNA content, and prognosis in human malignancy. J. Clin. Oncol., 5, 169&1703. MIX, M. C. 1975. Proliferative characteristics of atypical cells in native oysters (Ostrea lurida) from Yaquina Bay, Oregon. J. Invertebr. Pathol., 26, 289-298.

OPRANDY, J. J., AND CHANG, P. W. 1983. 5-bromodeoxyuridine induction of hematopoietic neoplasia and retrovirus activation in the soft-shell clam, Mya arenaria. J. Invertebr. Pathol., 42, 196-206. OPRANDY, J. J., CHANG, P. W., PRONOVOST, A. D., COOPER, K. R., BROWN, R. S., AND YATES, V. J. 1981. Isolation of a viral agent causing hematopoietic neoplasia in the soft-shell clam, Mya arenaria. J. Invertebr. Pathol.. 38, 45-51. PARDEE. 1989. G, events and regulation of cell proliferation. Science, 246, 603-608. PASTAN, I., AND WILLINGHAM, M. 1978. Cellular transformation and the ‘morphologic phenotype’ of transformed cells. Nature, 274, 645-650. PETERS, E. C. 1988. Recent investigations on the dis-

OF Mytilus

243

NEOPLASIA

seminated sarcomas of marine bivalve molluscs. Sot. Special Publ., 18, 74-92. PIERPAOLI, W., AND MESHORER, A. 1982. Host endocrine status mediates oncogenesis: Leukemia virusinduced carcinomas and reticulum cell sarcomas in acyclic or normal mice. Eur. J. Cancer Clin. Oncol., 18, 1181-1185. RABINOVITCH, P. S., KUBBIES, M., CHEN, Y. C., SCHINDLER, D., AND HOEHN, H. 1988. BrdUHoechst flow cytometry: A unique tool for quantitative cell cycle analysis. Exp. Cell Res.. 174, 309Am. Fish.

318.

ROBBINS, S. L., AND KUMAR, V. 1987. “Basic Pathology,” 4th ed. Saunders, Philadelphia. SHAPIRO, H. M. 1988. “Practical Flow Cytometry,” 2nd ed. A. R. Liss, New York. SMOLOWITZ, R. M., MIOSKY, D., AND REINISCH, C. L. 1989. Ontogeny of leukemic cells of the soft shell clam. J. Invertebr. Pathol., 53, 41-51. WITTKE, M., AND RENWRANTZ, L. 1984. Quantitication of cytotoxic hemocytes of Mytilus edulis using a cytotoxicity assay in agar. J. Invertebr. Pathol., 43, 248-253.

YEVICH, P. P., AND BARSZCZ, C. A. 1977. Neoplasia in soft-shell clams (Mya arenaria), collected from oil-impacted sites. Ann. N. Y. Acad. Sci.. 298, 409426.

ZAR, J. H. 1984. “Biostatistical Prentice-Hall, New Jersey.

Analysis,”

2nd. ed.