Int. J. Cancer: 23, 494-503 (1979)

ESTABLISHMENT A N D CHARACTERIZATION OF A HUMAN NULL-CELL LY MPHOBLASTIC LYMPHOMA CELL LINE (K-LL-3) Stephen D. SMITH and David ROSEN University of Kansas Medical Center, Department of Pediatrics, Kansas City, Kansas 66 103, USA

A new cell line (K-LL-3) with exceptional characteristics was derived from atbone-marrow aspirate from a patient with poorly differentiated lymphoblastic lymphoma. The cell line was established and the cells maintained on agar feeders for over 1 year. The morphology and cytochemical staining of the primary tumor and the cell line were remarkably similar. The growth characteristics, chromosome pattern, cell-surface receptors and Epstein-Barr virus studies clearly differentiated this cell line from non-malignant lymphoblastoid cell lines. The assay technique used was able t o correlate in vitro colony growth with the patient’s clinical course. The cells grew on the agar as colonies rather than as a single-cell suspension, and individual colonies could be aspirated and successfully passaged. The K-LL-3 cells lacked cell-surface markers (cytoplasmic and surface immunoglobulins, Fc receptors, C, receptors, SRBC) except HL-A and thus were classified as null cells. These cells were EBNA (Epstein-Barr virus-specific nuclear antigen) negative and had a pseudodiploid 46XY karyotype.

When human Burkitt lymphoma cells are cultured in vitro, they often give rise to a permanent cell line. The success rate for establishing these cell lines has been as high as 50% in some investigators’ laboratory (Nilsson, 1977). However, it has been difficult to establish permanent cell lines from non-Burkitt lymphoma. Two main problems exist in growing non-Burkitt lymphoma cells in vitro. First, lymphoblastoid cells derived from B lymphocytes (containing Epstein-Barr virus) often overgrow the cancer cells in an in vitro environment. Thus a contaminating cell line may be established which originated from the tumor biopsy but which is non-malignant in nature. The second problem is that often the malignant lymphoma cells, for reasons not yet understood, simply will not grow with established culture techniques and indeed only a few Epstein-Barr virus-negative lymphoma cell lines have been documented (Nilsson, 1977). An agar assay system has recently been developed that resulted in growth of malignant colonyforming cells from bone-marrow aspirates from children with acute lymphoblastic leukemia (Smith et al., 1978). This technique has been expanded to allow growth of colonies of malignant lymphoma cells. This assay system is especially appealing for it permits early separation of malignant from non-malignant colonies. This paper reports the establishment and characteristics of a null-cell lymphoma cell line (K-LL-3). It is proposed that K-LL-3 is a malignant lymphoma cell line, for it has the same cytochemical and morphologic characteristics as the original tumor and it does not contain the Epstein-Barr virus.

MATERIAL A N D METHODS

Clinical data The K-LL-3 cell line was derived from a bonemarrow aspirate from a 13-year-old Caucasian male. This patient was diagnosed at the University of Kansas Medical Center (KUMC) as having a diffuse, poorly differentiated lymphoblastic lymphoma according to the classification of Rappaport (1966) and this diagnosis was subsequently confirmed by the pathologists at the National Cancer Institute (NCI). At the time of original diagnosis (6/16/76) this patient was found to have diffuse tumor involvement in the abdomen and pelvis but the bone marrow and spinal fluid examination showed no evidenc: of tumor infiltration. The patient was treated with a combination of radiation therapy and chemothzrapy and he experienced a complete remission for 8 months. In March 1977, the patient developed meningeal disease followed by progressive bone-marrow involvement (Table I). He responded poorly to further therapy and died 13 months after diagnosis. Culture methods Bone-marrow aspirates were performed in the routine manner, and the first aspirate was used for morphologic examination of the cells. A second aspirate of 1 ml was anticoagulated with preservativefree heparin (200 units) and these cells were cultured as described elsewhere (Smith et al., 1978). Basically, this technique consisted of culturing cells in a Petri dish containing two layers of agar: a feeder layer and an overlayer. The medium used was modified McCoy 5A (Pike and Robinson, 1970) supplemented with 15 % calf serum (K.C. Biological Company, Lenexa, Kansas). The feeder layer was prepared by drawing 30 ml of blood from healthy, fasting volunteers on the morning of the experiment. The blood was anticoagulated with preservative-free heparin and left to sediment at room temperature for 2 h. The WBC-rich plasma layer was aspirated and the number of nucleated cells counted in a hemocytometer. A mixture of modified McCoy’s 5A medium and agar (final agar concentration 0.5%) was prepared and the WBCs added to a final concentration of 1 x lo6 cells/ml. One ml of this mixture was plated onto 35-mm tissue culture dishes. The overlayer was prepared by centrifuging the bone-marrow cells as previously described (Smith Received: September 29, 1978 and in revised form December 29, 1978.

495

A HUMAN NULL-CELL LYMPHOMA LINE TABLE I RESULTS OF SEQUENTIAL BONE-MARROW CULTURES: CORRELATION OF I N VITRO GROWTH WITH CLINICAL COURSE Bonemarrow aspirate Date

In vitro culture

Clinical course Normal colonies myeloid-monocytic

4/29/77 6/01/77 7/01/77

Stable disease Progressive disease Progressive disease

M-1

Normal M-3 50% lymphoblasts M-3 70 % lymphoblasts

Malignant colonies lymphoblastic '

Cell line

85

20

K-LL-1

20

2

K-LL-2

0

3200

K-LL-3

' Results expressed as number of colonies grown. 5 x 10' cells/plate were used in each experiment. Each number is the mean of four Petri dishes.

et al., 1978) and resuspending them in medium. The cells were washed and mixed with agar to a final agar concentration of 0.3% and a cell concentration of 5 x lo5cells/ml. One ml of this mixture was then added to 1 x lo6 autologous RBC and plated on top of the previously prepared feeder layer. These dishes were then placed in a plexiglass chamber (Richter, 1973) and gassed with a 10% carbon dioxide, 7 % oxygen and 83% nitrogen mixture. The chamber was then sealed and the dishes incubated at 37" C for 3 weeks. At the end of 3 weeks, the plates were removed from the chamber and the number of colonies were counted. A colony was defined as a cell group containing more than 50 cells on day 21 of incubation. No mitogens, antibiotics or antifungal agents were added to the media. No 2-mercaptoethanol (2 ME) or other thiols were used in this culture system. Establishment of the cell line After 21 days of bone-marrow culture, individual colonies from the primary cultures were aspirated

with a Pasteur pipet and transferred to a fresh feeder. No effort was made to remove all the agar from the cells. Aspirated colonies were mixed with 1 ml of the overlay mixture in the same manner as the original bone-marrow specimen.

Maintenance of the cell line The established cell line (K-LL-3) was maintained on agar feeders in the plexiglass chamber or in an incubator gassed with 10% CO, in air. After 2 weeks of growth, 1 ml of fresh medium was added to the top of the overlayer. After 3 weeks of growth the individual colonies were loosened with a Pasteur pipet and the free cells aspirated and replated on fresh feeders. Growth characteristics The growth of the established cell line with various media, sera, and peripheral blood cells was compared (Table 11). In each experiment, the overlay contained 1 x lo4 K-LL-3 cells. The cell viability was checked by the trypan blue exclusion

TABLE 11 CLONING EFFICIENCY OF K-LL-3 CELLS WHEN PLATED AT A CELL CONCENTRATION OF I x 1 0 4 CELLSIDISH Culture conditions

Cloning etliciency

1 . Modified McCoy 5A medium 2. Modified McCoy 5A medium+ 5 % CS 3. Modified McCoy 5A medium+ 15 % CS 4. Modified McCoy 5A medium 15% CS (dialyzed) 5. Modified McCoy 5A medium+l5% FCS 6. Modified McCoy 5A medium+l5 % CS (heat-inactivated) 7. Modified McCoy 5A medium+l5% CS+lO% HP 8. Modified McCoy 5A medium+l5% CS+lO% HPf3 x lo6lymphocytes 9. Modified McCoy 5A medium 15 % CS +20 % HP 10. Modified McCoy 5A medium+15% CS+40% HP 1 1 . Modified McCoy 5A medium + 5 x 2ME 12. Modified McCoy 5A mediumfvit C (50 mg/l)+5% CS

0 0

5.9%

+

0

6.2 %

9.5 %

13.6% 15.5% 22.2 % 21.6%

+

0 0 ~~

' C.S., calf serum; F.C.S., fetal calf serum; H.P., human plasma; ZME, mercaptoethanol.

496

SMITH A N D ROSEN TABLE TI1 CYTOCHEMICAL PROFILE O F K-LL-3 CELLS A N D BONE-MARROW ASPIRATE

Cytochemical staining

Reference

K-LL-3 cells

PAS Oil red 0 Acid phosphatase Sudan black B Peroxidase a-naphthyl acetate esterase Naphthol AS-D chloracetate esterase

McManus ( I 946) Lillie ( I 948) Barka and Anderson ( I 962) Sheehan and Storey ( I 947) Kaplow ( I 975) Loffler (1961) Moloney et a / . (1960)

t t

’ Bone-marrow aspirate done 7/1/77. -

+, strongly positive reaction;

t

technique (Paul, 1975) before each experiment (usually >95 % viable) and all overlays were placed on top of freshly prepared feeders. The number of colonies formed were counted after 21 days of culture and the results of the experiments expressed as percentage cloning efficiency (No. of colonies formed/No. of cells plated x 100). Growth of cells in liquid K-LL-3 cells were added to tissue culture bottles at concentrations ranging from 1 lo5 to 1 x lo6 I

FIGURE 1 - Cytocentrifuge preparation of a colony o f K-LL-3 cells. Most cells show a high nuclear-cyto-

plasmic ratio and the nuclei demonstrate a fine chromatin pattern with well-defined nucleoli. There is a mitotic figure in mid-field. Wright’s stain ( x 380).

+

1

Tumor cells in bone~

marrow aspirate

++ NT

t

I ~

~

-

NT

-

-

-

-

t , positive reaction; ---, negative reaction; NT. not tested

cells/ml. These cells were incubated at 37” C in a humidified incubator gassed with 10% CO, in air or gassed and resealed in the plexiglass chamber. One-third of the culture medium was renewed twice a week. Morphological methods Light microscopy. K-LL-3 cells were removed from the agar in two ways for staining with Wright’s stain. Individual colonies were aspirated with a Pasteur pipet, smeared between two cover-slips and then stained. Cells were also stained after cells in liquid were cytocentrifuged onto glass slides. Both preparations gave identical results. Transmission electron microscopy ( T E M ). The K-LL-3 colonies were individually aspirated from the agar by a Pasteur pipet and washed twice with media. The cells were fixed in 2% glutaraldehyde, post-fixed in osmium tetroxide and embedded in Luft’s epon mixture. The blocks were sectioned on an LKB ultratome I11 and the sections were stained with uranyl acetate and lead citrate. The specimens were viewed and photographed in an RCA E M U 3G transmission electron microscope. Scanning electron microscopy ( S E M ) . The K-LL-3 colonies were aspirated and washed as described for TEM (above) and then layered onto polylysinecoated glass slides. The cells were kept moist, fixed with 2% glutaraldehyde, and then critical-pointdried using Freon 13 in a Bomar SPC-I500 critical point drier. The cells were then gold-coated in a Tecnics Hummer coater and viewed and photographed in a JEOL JSM-35 electron microscope.

Cell size measurements Cells from actively growing colonies were removed from agar as previously described and suspended in medium. Cell viability was measured by the trypan blue exclusion technique (Paul, 1975) and cell size determinations were done on cell populations with more than 98 % viability. Cell size determinations were made on 500 cells on two separate occasions using a Zeiss optical micrometer (Needham, 1968). Cytorhemistvy The K-LL-3 cell line and the patient bonemarrow lymphoblasts were analysed using a variety

A H U M A N NULL-CELL LYMPHOMA LINE

of cytochemical stains. All staining was done in duplicate, and the staining procedures are referenced in Table 111. Surface receprors T, B, and pre-B assays were performed on colony cells that were removed from the agar by a Pasteur pipet and incubated in medium overnight to enhance separation of the agar from the cells. T cells were identified by their ability to form rosettes with sheep erythrocytes (SRBC) (Froland, 1972). The colony cells were assayed for the presence of cytoplasmic immunoglobulin as described previously (Gathing et af., 1977). Briefly, cells were cytocentrifuged (Shandon-Southern Industries, Sewickley, Pa.) onto glass slides, fixed at ambient temperature for 10 minutes in acetone and stained by direct

FIGURE 2 - Scanning electron micrograph of representative K-LL-3 cells. Most of these cells had a smooth surface with only small folds and projections.

x 6,600 (40" tilt).

immunofluorescence with fluorescein isothiocyanate conjugated with anti-IgM, anti-IgA and anti-IgG (Hyland Laboratories, Costa Mesa, Calif.) at a dilution of 1:s. Membrane IgM and IgD were assayed (Winchester et al., 1975) by direct immunofluorescence with aggregate-free (centrifuged at 140,OOOg for 90 min) fluorescein conjugated antihuman IgM and anti-human IgD. These reagents are routinely used to detect B cells in mononuclear cell populations derived from human peripheral blood. Fc receptors were detected by a standard EA (erythrocyte antibody) assay. Briefly, sheep erythrocytes were coated with a 1:3,000 dilution of rabbit anti-sheep erythrocyte serum. Test cells (lo6 cells in 0.1 ml) were mixed with EA (0.1 ml of 0.5% cell suspension) for 5 min at 37" C, sedimented at 50g for 5 min, resuspended and assayed for percentage rosettes. Complement receptors

497

are detected by a modification of the above assay wherein sheep erythrocytes are sensitized with a 1 :6,000dilution of the rabbit anti-sheep erythrocyte serum and fresh human serum. The fresh human serum is titrated to produce slight hemolysis. This EAC assay is performed as described above for the EA assay. The sensitivity of each of these rosette assays has been established using cells from a variety of sources, the most critical being peripheral blood mononuclear cells from normal human as well as patients with chronic lymphocytic leukemias (B lymphocytes). immunoglobulin secretion Secretion of IgG by the K-LL-3 cells was tested by an IgG immunofluorescence assay (Bio-Rad Laboratories, Richmond, Calif.). Optimally growing

FIGURE 3 - Transmission electron micrograph of a representative K-LL-3 cell. Nucleus is lobated and cytoplasm shows polyribosomes, mitochondria and vacuoles. ( x 9,350).

K-LL-3 cells were cultured in serum-free medium for 24 h and then the supernatant was collected and concentrated 100-fold. This supernatant (plus samples of normal serum and cerebrospinal fluid as controls) was mixed with antibody to human IgG that was covalently coupled to small hydrophilic beads. This mixture was incubated at 37" C for 90 min and then fluorescence-labelled monospecific (IgG) antiserum was added and incubated for 60 min. The samples were washed twice by centrifugation and resuspended in buffer and the relative fluorescence determined in a fluorometer. Standard curves were calculated from IgG standards supplied in the assay and the lower sensitivity limit was 100 ng/ml.

498

SMITH AND ROSEN

-"--~---

w---G--

HL-A typing

K-LL-3 cells were removed from the agar and HL-A typed by the microtoxicity technique (Manual of Tissue Typing Technique, 1977). Peripheral blood samples were submitted from the patient for HL-A typing on two separate occasions. Chromosomes

K-LL-3 cells were harvested and suspended in medium at a concentration of 1 X lo6 cells/ml. Trypsin modification of Giemsa banding was accomplished on the growing cells without PHA stimulation (Patil et al., 1971). The cell line was initially tested after 4 months of growth and then retested during the 16th month of growth. Tests for Epstein-Barr virus (EBV)

EBV studies were performed on the K-LL-3 cells on two separate occasions: after 5 months and after 12 months of continuous growth. K-LL-3 cells were harvested and washed twice with medium, centrifuged for 5 min, and resuspended in 0.5 ml fetal calf serum. Smears were prepared on glass cover-slips and allowed to dry at room temperature for 1 h. The cells were then fixed in a 1:l mixture of cold methyl alcohol and acetone for 5 min. The cover-slips were then shipped to Dr. Werner Henle, Division of Virology, The Children's Hospital of Philadelphia, who tested the cells for EBVassociated nuclear antigen (EBNA). Peripheral blood from the patient was drawn both at diagnosis and 2 months later and sent to the Center for Disease Control (Atlanta, Georgia) for determi-

FIGURE 4 - Karyotype of a cell from the K-LL-3 line showing loss of both chromosomes 8 and one cl.ron ?some 2. There is also one eirm chimosome 5 , 13 and 19 resulting in a pseudodiploid 46XY karyotype.

nation of Epstein-Barr virus titers. The blood samples were tested for EBV-antibodies directed against viral capsid antigen (VCA). RESULTS

The pathologic specimen of this patient's abdominal tumor was classified as diffuse undifferentiated lymphoblastic lymphoma at both KUMC and the NCI. This tumor responded initially t o chemotherapy and then metastasized to the bone marrow. Table I shows the correlation between in vitro studies and the clinical status on three successive bone-marrow examinations. Identification of colony types

Normal myeloid and monocytic colonies grew side by side with the malignant lymphoblastic colonies (Table I). After 21 days of culture, individual colonies were aspirated and classified on the basis of their cytochemical staining pattern. The cells of the myeloid colonies grew close together on the agar and had only a loose rim of migrating cells. These colonies were mostly composed of myelocytes and metamyelocytes, the cells were peroxidasepositive and the cytoplasm stained faintly and diffusely with PAS. The cells of the monocytic colonies grew in a loose arrangement on the agar and these colonies consisted of monocytes that were peroxidase-positive and PAS-negative. At the end of 21 days of culture, both the myeloid and monocytic colonies appeared to be degenerating, for many of the cells had irregular borders

499

A HUMAN NULL-CELL LYMPHOMA LINE

and distorted shapes. The cells of the lyrnphoblastic colonies grew tightly together and these cells showed little migration. These colony cells appeared as lymphoblasts on Wright’s stain (see “Morphology of cell line”) and they were PASpositive and peroxidase-negative. At the end of 21 days of culture, the lymphoblastic colony cells were round with a distinct border and some possessed cytoplasmic vacuoles that could be seen on high magnification (1,000 x).

Establishment of the cell line Lymphoblastic colonies from the primary culture regularly continued to grow when transferred onto a fresh feeder after 21 days of growth. Myeloid and monocytic colonies were not successfully p-issaged. Continued growth of the transferred co onies could be identified as early as 3 days after transfer. The initial bone-marrow cultures showed that the lymphoblastic colonies grew better in the

plexiglass box (Po2 = 40 mmHg) compared to an incubator gassed with 10% CO, in air (Po, = 160 mmHg) (Table V). However, passaged cells grew equally well in both environments when 1 x lo4 cells/ml were plated. The cell lines were maintained in the standard incubator gassed with 10% CO, in air because it is much easier to use than the plexiglass box. Cell lines were established from 3 bone-marrow samples (Table I). Cell line K-LL-1 and K-LL-2 were lost due to fungal contamination after about 10 passages. Cell line K-LL-3 has been cultured for over 1 year and more than 70 passages have been successfully performed.

Growth characteristics of the established line K-LL-3 cells grow rapidly in the agar assay system. The doubling time of the passaged cells is 24-36 h after a latent period of about 48 h. For sustained growth, the cells are still (after 13 months of culture) dependent upon a feeder layer and

TABLE IV COMPARISON OF K-LG3 CELLS WITH PREVIOUSLY DESCRIBED LYMPHOCYTIC LYMPHOMA A N D LYMPHOBLASTOID CELL LINES Cell line K-LL-3

Reference

Cell type Growth pattern In suspension Feeder dependency Initial latent period Surface characteristics Surface immunoglobulins Fc receptor C, receptor Pre-B SRBC Immunoglobulin secretion Karyotype

EBV genome NT, not tested.

U698-M

U715-M

Nilsson and Sundstrom

Nilsson and Sundstrom

sn2

D.G 75

Ben-Bassat

Schneider and Schwenk

LCL

(1974)

(1 974)

(1977)

Nilsson and Ponten (1 975) Nilsson (1 977)

Nilsson

Nilsson

(1977)

(I 977)

Lymphocytic lymphoma

Lymphocytic lymphoma

Lymphocytic lymphoma

Lyniphocy tic lymphoma

Schneider al. (1977) Lymphocyt ic lymphoma

Lymphoblastoid

None Absolute

Single-cell None

Clumps None

Small clumps Single cells None None

Large clumps None

3 days

6 weeks

6 weeks

8 weeks

9 weeks

6 weeks

Negative Negative

Positive Positive

Positive Positive

Positive Negative

Negative NT

Negative Negative Negative None

Negative NT Negative None

Positive NT Negative None

Negative NT Negative NT

NT NT Negative NT

Positive Negative to weakly oositive Positive NT Negative Yes

Pseudodiploid (missing C and extra B)

Aneuploid (trisomy 7)

Aneuploid (trisomy 7)

Negative

Negative

Negative

Diploid with chromosome 14 asymmetry and some tetraploidy Negative

(1977)

el

Diploid with D14q+

Normal diploid

Negative

Positive

500

SMITH AND ROSEN

complete medium supplemented with calf serum (Table 11). Human plasma enhances the colony growth, and a cloning efficiency as high as 22% has been reached (Table 11). Neither 2-mercaptoethanol nor vitamin C could replace calf serum in this assay. When colonies from the primary culture were passaged into bottles containing medium alone, no growth resulted. Colonies were also transferred onto feeder layers using an overlay that contained no agar. Often the lymphoblasts would then grow out of the agar and become loose cells in the medium. These cells did not attach to the feeder layer and could easily be transferred. All cells transferred to plates without feeders showed no growth. Morphology of cell line Light microscope. The primary and passaged lymphoblastic colony cells had identical morphology. Wright’s stain showed large, undifferentiated cells that appeared to be lymphoblasts (Fig. 1). The nuclear-cytoplasmic ratio was high with only slight variation from cell to cell. The nucleus was round and only a few cells had nuclear clefts. There were usually one o r two well-defined nucleoli

dant polyribosonies. Occasional glycogen and sparse endoplasmic reticulum could be seen. Chromosomes Both of the chromosome assays on the K-LL-3 cells showed a basic pseudodiploid male karyotype. In the initial testing (after 4 months of growth) six of the eight spreads showed a 46XY karyotype with one extra B-group chromosome and one missing C-group chromosome. One spread showed a 45XY pattern with one missing C chromosome and another showed 47XY karyotype with one missing C- and two extra B-group chromosomes. Retesting after 16 months of growth showed some clonal evolution of the cells in culture but the basic pseudodiploid male karyotype pattern persisted. Twenty-six spreads were analyzed and 12 showed a pseudodiploid 46XY pattern. Abnormalities that occurred fairly consistently were loss of an A-group and a C-group chromosome with extra B- and F-group chromosomes. Some cells had an additional D-, an additional A-, and an additional E-, F- or G-group chromosome. Some cells showed a loss of one A- and two Cgroup chromosomes combined with extra B, D and F chromosomes (Fig. 4).

TABLE V INFLUENCE OF OXYGEN CONCENTRATION ON K-LL-3 CELLS Number of malignant lymphoblastic colonies grown Cell sample

Plexiglas box (7 ”/. oxygen, 10 % C o d

Primary bone-marrow culture 4/29/77 5 x 1O5 bone-marrow cells/plate Passaged K-LL-3 cells 1 x lo3 K-LL-3 cells/plate 5 x lo3 K-LL-3 cells/plate 1 x lo4 K-LL-3 cells/plate 5 x lo4 K-LL-3 cells/plate 1 x lo5 K-LL-3 cells/plate

per nucleus. The cytoplasm was moderately basophilic and some of the cells had cytoplasmic vacuoles. There were no cytoplasmic granules and only a rare cytoplasmic projection. The cells were heterogenous in size, the mean size being 12.3 pm with a range of 7.5-17.8 pm (1.2 SD). These cells looked identical to the malignant lymphoblasts that were present on smears made directly from the bonemarrow aspirate. Ultrastructure. With SEM (Fig. 2), the majority of the K-LL-3 cells had a smooth surface with small folds and projections. These cells demonstrated attachment with microfilamentous processes extending from the cell onto the glass slide. TEM showed that the cells had lobated nuclei (Fig. 3) and a well-developed nucleolus. The cytoplasm contained vacuoles, mitochrondia and abun-

Standard incubator ( 2 0 % oxygen, 10% CO,)

20

3

0

0 0 816 2,140 3,897

326 778 1,897 4,342

Epstein-Barr virus studies K-LL-3 cells were studied by Dr. W. Henle on two occasions: the first after 5 months of growth (12/77) and the second after 12 months of growth (7/78). On both occasions, the Epstein-Barr nuclear antigen (EBNA) tests showed no reactions in the K-LL-3 cells. Serum from the patient was not available at the time of the cell line testing (12/77); however, serial Epstein-Barr virus titers on the patient’s peripheral blood were less than 1 :I0 when performed in June and August of 1976. Cell-surface markers and HL-A Lymphoblasts from the bone-marrow aspirates lacked specific cell-surface markers (surface inimunoglobulins, Fc receptors, C , receptors and SRBC) except HL-A, and the patient’s disease was classi-

A HUMAN NULL-CELL LYMPHOMA LINE

fied as a null-cell lymphoma. The K-LL-3 cells also lacked these markers (Table IV) and the tumor cells did not secrete IgG. The HL-A type of K-LL-3 (the first cell line established in our laboratory) was: A-2, A-11, B-37, and B-5 or B-15. There was cross-reactivity at the B-5 and B-15 locus, making precise identification impossible. HL-A typing of the patient’s blood was unsuccessful on two occasions because the white cell count was too low. DISCUSSION

K-LL-3, a null-cell lymphoma cell line, has been established and passaged on an agar assay system for over 1 year. This assay system has permitted the growth of malignant colony-forming cells from bone-marrow aspirates from children with acute lymphoblastic leukemia (Smith ef al., 1978) and it may have some advantages over growth of malignant cells in liquid (Metcalf, 1977). This is the first lymphoma cell line established by this assay and the K-LL-3 cell line is one of only a few which have been established from human lymphoblastic lymphoma (Nilsson, 1977). The K-LL-3 cells have several distinctive features in that growth of the primary tumor was observed 3 days after plating, the cells required a feeder for growth even after 1 year of passage, and the cells were characterized as null-cells (lacking cell-surface markers except HL-A). The original bone-marrow aspirates were incubated in a hypoxic environment (Po2 = 40 mmHg) because this environment has been shown to enhance growth of malignant colony-forming cells in acute lymphoblastic leukemia (Smith et al., 1978). The results of K-LL-3 cells incubated in two different oxygen environments are shown in Table V. In the primary bone-marrow culture, the plates incubated in a 7 % oxygen environment showed slightly enhanced growth compared to cells incubated in 20% oxygen. This stimulating effect of hypoxia could also be demonstrated when low cell concentrations of established K-LL-3 cells (5 x lo3 cells/plate) were cultured, but this effect was lost at higher cell concentration ( 1 x lo4 cells/plate). The beneficial effect of hypoxia appears to be replaced by cell-cell interactions at higher cell concentrations. This enhancement of growth at low cell concentrations is important, for large volumes of bone-marrow malignancies cannot be obtained. Also, one of the main problems in working with bone-marrow malignancies is that 1 ml of bone-marrow cells represent less than 1 millionth of the patient’s total tumor burden (Frei and Freireich, 1965). Thus the original inoculum may contain only a few malignant cells that possess the potential to grow on agar. In the agar assay system used, the donor’s bonemarrow sample grew malignant colonies side by side with both normal myeloid and normal monocytic colonies. After a few weeks of growth, the colonies were harvested and transferred to new feeders. After the colonies had been passaged, only the malignant lymphoblastic colonies grew, for normal myeloid and monocytic colonies will

50 1

not grow once passaged (Metcalf, 1977). Following a few additional weeks of growth, the lymphoblastic colonies were often large enough to be re-aspirated, manipulated free of agar and transferred to a fresh feeder. Thus a cell population containing only lymphoblasts was obtained during the first month of primary culture growth. In this way, slowly growing fibroblasts and B lymphocytes containing Epstein-Barr virus were eliminated from the assay before they had a chance to grow. These latter cells usually have a latent period of 6-8 weeks and if not removed from the assay they can overgrow the malignant colonies (Nilsson and Ponten, 1975). In the early stages of the primary culture, only intact colonies could be passaged, and if the colony broke down into a single-cell suspension, no growth occurred. Later, when the colony contained >1 x lo4 cells, singleeelf suspensions of 1 x lo4 cells/ml readily grew with a cloning efficiency approaching 8 % (Table V). All mitogens were excluded in this assay system to specifically avoid stimulation of nonmalignant cells that might be present in the bone-marrow aspirates. 2-Mercaptoethanol and other thiols were excluded because they have been shown to have a co-mitogenic effect (with agar) on mouse B lymphocytes (Kincade et al., 1976). The K-LL-3 cells did not grow well in liquid and stopped growing after a few weeks unless reseeded onto a feeder layer. This absolute dependency on a feeder layer (for over one year) is unusual, for malignant lymphoma cells usually grow independent of feeders after a few months in culture (Nilsson, 1977). The cells were also dependent on calf serum for growth and vitamin C and 2 ME could not replace this requirement (Table 11). Human plasma augmented the growth of the K-LL-3 cells, and a cloning efficiency as high as 22 % was consistently obtained (Table 11). The proposal that this cell line originated from the donor’s tumor cell population is supported by two lines of evidence. First, the K-LL-3 cells had the same morphology, cytochemical staining and cell-surface markers as the malignant bonemarrow cells (Table 111, LV). Non-malignant lymphoblastoid cell lines (LCL) often grow from malignant tumor biopsies (Nilsson, 1977), but the K-LL-3 cells differed from lymphoblastoid cells in many basic characteristics. The LCL is characteristically Epstein-Barr virus (EBV)-positivz, possesses a diploid chromosome number, has surface immunoglobulins and secretes immunoglobulins (Nilsson and Ponten, 1975; Nilsson, 1977). The K-LL-3 cells are EBV-negative, have a pseudodiploid chromosome number, have no surface immunoglobulins and do not secrete immunoglobulins (Table IV). The other cell lines have a latent period of 6-8 weeks while growth of the primary culture of the K-LL-3 cells was evident after 3 days. This may be a function of the agar assay system used because early growth of clusters of 6-10 cells can easily be seen on agar but not in a cell suspension. This K-LL-3 cell line lacked surface markers (cytoplasmic and surface immunoglobulins, Fc receptor, C, receptors, SRBC) except

502

SMITH AND ROSEN

HL-A and thus has been classified as a null-cell line (Table IV). Of the other lymphoma cell lines grown, only the SH-2 cell line possessed similar cell-surface characteristics, but it did not survive long enough (grew for