Interculture variation and evolution of B lineage lymphocytes in long-term bone marrow culture

Eur. J. Immunol. 1986.16: 779-787 Interculture variation in long-term bone marrow cultures 779

Interculture variation and evolution of B lineage Pamela L. Witte, Paul W. Kincade and Vaclav V6tviiSka’

The Oklahoma Medical Research Foundation, Oklahoma City

lymphocytes in long-term bone marrow culture*

A recently described long-term culture system offers a unique experimental approach for dissecting regulatory mechanisms that control the developmental progression of B-lineage lymphocytes. Lymphoid cells, including B cells and their precursors, can be maintained for prolonged periods in strict dependence on a layer of adherent cells. However, before this system can yield to interpretable manipulation, much informa- tion is needed as to the identity and temporal phenotypic stability of both lymphoid and nonlymphoid cells. The findings reported here provide answers to some of those important questions. Successful establishment of lymphoid cells in culture was extra- ordinarily dependent on the batch of fetal calf serum used in the medium, and some undesirable serum lots supported cultures that were virtually all myeloid. With standardized culture conditions, various populations of lymphoid cells were identified on the basis of B-lineage differentiation markers and culture to culture variation was assessed. Lymphocytes that were firmly attached to the adherent cells were carefully compared to nonadherent lymphocytes in terms of cycle status, phenotype, size, and transferrin receptor expression. They were essentially identical in all of these respects and a partitioning of proliferating cells and their progeny in the cultures was therefore not apparent. It is also noteworthy that although a high mitotic rate was maintained, a majority of the cells were small lymphocytes. The outgrowth of identifiable B-lineage cells (detected with monoclonal 14.8 antibodies) in replicate cultures was initially similar, but the extent of interculture variation increased dramatically during the period 4-6 weeks after initiation of culture. Replicate cultures established from the same marrow cell pool often differed as much as 20-fold in numbers of 14.8-positive cells. After this time, the composition of individual cultures evolved much more gradually, and numbers of B cells and pre-B cells remained relatively constant. This indicates that subsets of lymphocytes become established in each culture dish during a discrete phase. At least two types of supporting adherent cells predominated in these cultures: typical macrophages and very large, nonphagocytic cells resembling adventitial reticular cells. The latter included subpopulations resolved on the basis of alkaline phosphatase content. In contrast to the lymphoid populations, proportions of these adherent cell types were relatively invariant among replicate cultures. The dramatic expansion or contraction of particular B-lineage populations which occurred prior to 6 weeks of culture did not appear to be reflected in the composition of the adherent layers on which they depended.

1 Introduction

Little is known about the mechanisms that control the sequence of events through which progenitor cells give rise to B lymphocytes. Unlike T lymphocytes, mammalian B cells arise within the hemopoietic compartment of the bone marrow (BM) or fetal liver, surrounded by a multitude of other cell types and without apparent anatomical sequestering [l, 21. Microenvironmental “niches”, similar to those which are

[5432]

* This work was supported by grants from the National Institutes of Health AI-200069 and AI-19884 and by a Damon Runyon-Walter Winchell Cancer Fund Fellowship, DRG-739 (to P. L. W.).

A Current address: Department of Immunology, Institute of Mi- crobiology, Prague, Czechoslovakia.

Correspondence: Pamela Witte, Oklahoma Medical Research Foun- dation, 825 N.E. 13th Street, Oklahoma City, OK 73104, USA

Abbreviations: LTBMC-B: Long-term bone marrow culture of B- lineage cells FCS: Fetal calf serum mAb: Monoclonal antibody Ig: Immunoglobulin FALS: Forward angle light scatter sIgM: Sur- face IgM SRBC: Sheep red blood cells cp: Cytoplasmic heavy chains of IgM

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1986

thought to influence myeloid and erythroid cell differentiation [3-51, are essentially unknown for developing B lymphocytes. Yet, evidence from several experimental approaches indicates that the progenitors of B-lineage lymphocytes depend on extrinsic factors for appropriate differentiation. First, Sephadex G-10-adherent BM cells facilitate the ability of pre- B cells to become B cells in short-term culture systems [6]. Second, soluble moieties can be detected in the serum or urine of mice and patients with dysregulated lymphopoiesis and hemopoiesis [7, 81. Third, a recently introduced system for the long-term propagation of B-cell progenitors [9] is similar to the long-term BM culture (LTBMC) system devised by Dexter and colleagues [lo] which supports multipotential stem cells that give rise to myeloid lineage cells. Numerous studies with Dexter-type cultures have shown that a necessary communica- tion exists between hemopoietic precursors and microenvironmental cells [ l l] . Similarly, unless virally transformed, murine B-cell precursors are strictly dependent on nonlymphoid, adherent BM cells or factors produced by these cells for continued maintenance in long-term cultures [9].

Because comparable association between adherent cells and myeloid precursors in Dexter-type cultures have been noted between developing myeloid cells and BM stromal cells in situ

0014-2980/86/0707-0779$02.50/0

780 P. L. Witte, P. W. Kincade and V. VktviEka Eur. J. Immunol. 1986.16: 779-787

[4, 51, long-term culture systems may reflect many aspects of normal hemopoietic regulation. Long-term cultures of developing B lymphocytes, selectively maintained by adherent regulatory cells, thus offer a unique approach to understand- ing mechanisms that control the progression of B-lineage cells through their developmental pathway. We need to know more about the minimal requirements of lymphoid cell growth, we must determine if particular cell populations are maintained through self renewal alone or if normal differentiation events also occur, and the role of adherent “microenvironmental” elements in these processes should be analyzed in detail.

In this comprehensive study, we describe striking interculture variability of B-lineage cell phenotype and the apparent evolution of discernable early B cell subpopulations over time within individual long-term BM cultures. This variability of cultured lymphoid cells is in contrast to the minimal interculture disparity of the supporting adherent cells.

2 Materials and methods

2.1 LTBMC-B

2.1.1 Preparation of LTBMC-B

BALBkJ and NZB/BinJ female mice, purchased from Jackson Labs. (Bar Harbor, ME), were killed between 3 and 4 weeks of age by cervical dislocation. LTBMC were prepared by the method of Whitlock et al. [12]. Briefly, femoral marrow from 20-25 mice was pooled and dispersed into a single cell suspension. For lymphoid cell culture, 12 million nucleated cells (lo6/ ml) per 100-mm culture dish (#25020 Corning Glass Works, Corning, NY) were incubated at 37°C in 6.0% COz. The medium consisted of RPMI 1640, 5% fetal calf serum (FCS), 5 X M 2-mercaptoethanol (2-ME), 2 mM L-glutamine (Gibco, Grand Island, NY), 100 unitsiml penicillin and 100 pg/ ml streptomycin (Gibco). As a preventative measure against contamination, the culture dishes were wrapped loosely in Saran wrap, which allowed full gaseous exchange. The cultures were maintained by replacing three-quarters of the spent medium every 7 days. At midweek, 4 ml of fresh medium was added to each culture. At each step, care was taken to disrupt the cellular architecture as little as possible.

2.1.2 FCS

Nine serum lots from 4 distributors were screened for their efficiency in establishing lymphoid cell growth from weaning mouse BM. These included: lots 29101664 and 29101653 from Flow Labs (MacLean, VA); lots 100428 and 110451 from Hy- clone (Logan, UT); lots 4010167,311705,331704, and 310642 from Irvine Scientific (Santa Ana, CA); and lot 36k8141 from Gi bco .

2.1.3 Harvest of LTBMC-B

Lymphoid cells were harvested at the end of the 7-day feeding cycles. Suspended and loosely adherent lymphoid cells from LTBMC-B were harvested by gently swirling and flushing to suspend the cells. Lymphoid cells collected in this way are referred to as “nonadherent”. Clusters or foci of lymphoid cells associated with nonlymphoid adherent cells appeared to

be firmly attached and undisturbed by this manipulation. These “stromal-associated” lymphoid cells were mechanically dislodged by vigorously shaking the culture dishes on a Ther- molyne Maxi Mix shaker (Sybron, Dubuque, IA). Approxi- mately 75% or more of the stromal-associated cells could be harvested with minimal contamination by adherent cells. Alternatively, most stromal-associated cells were detached by incubation with 5 ml of 0.02% EDTA (w/v, Sigma) in Ca2+/ Mg2+-free Hanks’ balanced salt solution (HBSS; Gibco). The harvested cells were immediately centrifuged and resuspended in fresh RPMI 1640 plus HEPES and 5% FCS (heat inactivated, 56”C, 30 min) to lo6 viable cells/ml. The adherent cells from LTBMC-B were recovered by enzyme treatment and gentle scraping. We found that mechanical disruption alone yielded very low viability and that collagenase digestion (5 different preparations from CooperlWorthington, Freehold, NJ, were tested; data not shown) released at most only ode- third of the adherent cells. Therefore, we chose a 5-min treatment with 0.25% trypsin (wh) in the presence of 0.02% EDTA to disperse the adherent cells. First, stromal-associated lymphoid cells were removed with EDTA in HBSS, and the adherent cells were then washed with Ca2+/Mg2’-free HBSS (37 “C) before enzyme treatment. Five ml of trypsin/EDTA was added to each 100-mm dish and incubated at 37°C. The reaction was stopped with cold HBSS plus 10% FCS. Cells that remained attached were released by very gentle scraping with a wide silicone rubber policeman. The cells were washed twice and resuspended in Iscove’s medium (Gibco) plus 10% FCS, antibiotics and 2-ME. In these experiments, the viability was always > 90% immediately after treatment, and the cells appeared to remain healthy as most reattached and spread overnight. The efficiency of this procedure varied with the source of trypsin; a prepared solution purchased from KC Biologicals (Lenexa, KS) and low-contaminant, crystalline trypsin from Cooper/Worthington were used in these studies.

2.2 Phenotype of B-lineage lymphoid cells

The B lymphocyte lineage phenotype of lymphoid cells recovered from individual cultures was characterized as follows. All cells that would most likely belong to this differentiation pathway were identified by surface staining with 14.8 monoclonal antibody (mAb) [13] by flow cytometry or fluorescence microscopy. The latter was particularly useful during the first 4 weeks because of problems encountered with dead cells and debris present in the nonadherent fraction. Immediate precursors of B cells (pre-B) cells were detected by staining of fixed cytocentrifuged preparations for the presence of the p heavy chain of IgM (cp) along with simultaneous exclusion of surface IgM (s1gM)-bearing B cells.

2.3 Immunofluorescence analysis

Rat anti-mouse mAb to B-lineage cell antigens have been previously characterized by this laboratory [13-151. mAb 14.8 was used here as 1 0 ~ concentrated hybridoma culture supernatant. Affinity-purified, fluorescein-conjugated mouse anti- rat immunoglobulin (Ig) was prepared as described [15]. mAb 331.12 (anti-IgM), fluorescein conjugate, was from Becton Dickinson (Mountain View, CA). Rat anti-mouse mAb R17 217.1.3 [16] to the transferrin receptor was the generous gift of Dr. I. Trowbridge. Rat anti-mouse T200 mAb [17] was used as an ammonium sulfate precipitate from culture supernatant


(clone MlB9.18.7 from the American Type Culture Collec- tion, Rockville, MD). All antibodies were used at previously determined saturating conditions.

Cell surface staining was accomplished in 96-wel1, round-bot- tom polystyrene microtiter plates (Dynatech, Alexandria, VA) to screen small numbers of cells. Two hundred thousand cells per well were suspended in 10 pl of primary antibody for 15 min on ice. The plates were shaken once during each incubation. The cells then were washed once in 200 pl RPMI 1640 plus HEPES and 5% heat-inactivated FCS before suspending in the secondary antibody. The cells were washed twice more and analyzed by fluorescence microscopy or by flow cytometry using a Coulter Epics V (Hialeah, FL). Low forward angle light scatter (FALS) was used to exclude a majority of nonviable cells and debris. These gate settings corresponded to small and medium cells in normal BM. In several initial experiments, propidium iodide (Sigma) was used to directly identify nonviable cells; however, the viability of freshly harvested lymphoid cells was always > 90% and this step was omitted in later experiments. For each sample, at least 5 x lo3 viable nucleated cells were evaluated with respect to log-integrated green fluorescence and FALS. The data presented represent the portion of positive cells after channel by channel subtrac- tion of negative control fluorescence followed by calculation of the percentage of cells falling in channels beyond 36 (scale 1-256 channels). In control experiments, the range of 14.8' cells varied less than 3.0% between replicate wells containing normal BALB/c BM.

Pre-B cells were distinguished by the presence of cy in cells which lacked sIgM. Cells stained with mAb 331.12 were cytocentrifuged, fixed in ethanol-acetic acid, and counter- stained with goat anti-mouse p conjugated with rhodamine (Southern Biotechnology, Birmingham, AL). The immunofluorescence of at least 200 cells was scored using a photomi- croscope I11 (Carl Zeiss, Thornwood, NY) as described previously [ 181.

2.4 Cell cycle analysis

One to three million freshly harvested lymphoid cells were washed in saline G (Ca2+/Mg2+-free phosphate-buffered saline plus 6.1 mM glucose) and fixed overnight in 70% ethanol diluted in saline G. The fixed cells were subsequently washed, treated with 1000 units of RNAse (Cooper/Worthington) for 30 min at 37 "C and finally stained with 36 yg propidium iodide in 1 ml saline G for 20 min (protocol kindly provided by Dr. E. Ravache, Albany Medical College). Ten or twenty thousand FALS-gated cells were analyzed by flow cytometry. GO + GI, S, and G2 + M stages of the cell cycle were calculated with the Coulter PARA 1 statistical program.

naphthol AS-MX phosphate and fast blue RR salt (Sigma, 85L-2). Cells to be tested were grown either in glass chamber slides (Miles Scientific, Naperville, IL) or directly cytocentrifuged onto glass slides.

2.5.2 Phagocytic and pinocytic assays

Adherent cells from LTBMC-B were detached by enzymatic treatment and allowed to reattach and spread in Iscove's medium (with 10% FCS, antibiotics, and 2-ME) onto No. 1 and 1/2 glass coverslips or in chamber slides for 48 h. Latex beads or fluoresceinated HEMA (a methacrylate copolymer [ 191) microspheres were used to assess general phagocytic ability. Phagocytosis, mediated by attachment of particles to the receptors for Fcy, was performed with sheep erythrocytes (SRBC) coated with the maximal, subagglutinating concentra- tions of anti-SRBC mouse mAb of the IgM, IgG2,, and lgGzb subclasses (hybridoma clones N-S.2.1, S-S.l, and N-S.8.1 were acquired from American Type Culture Collection). Fluoresceinated dextran (mol. wt. 70000, Sigma) was incubated with the cells for 2 h to determine the ability to rapidly pinocytose small molecules. At least 200 cells were scored on each slide by phase or fluorescence microscopy.

3 Results

3.1 Characteristics of cell outgrowth

Our preliminary experiments revealed an extraordinary serum requirement for successful growth of lymphoid cells. Cultures initiated with medium containing different FCS lots consisted predominantly of monocytes, predominantly of granulocytes, mixed populations, or, in the case of Hyclone lot #110451, virtually all lymphocytes [20].

In subsequent experiments, with lymphocyte-permissive FCS, pools of young (3-4 week) BALB/c or NZB BM cells from groups of 20 mice were used to initiate replicate cultures. These were harvested at intervals and carefully evaluated over a 12-14-week period. Cell numbers, morphologies, and phenotypes were compared. Less than 10% of the starting numbers of cells were present in the nonadherent cell fractions after one week (Fig. 1). Foci of lymphoid cells, associated with adherent cells, appeared around three weeks of culture and consistently expanded in number until a peak was reached (usually at 5-6 weeks). Cell numbers then stabilized at a slightly lower level. Percentages of nonadherent cells consid- ered to be lymphoid by morphology always exceeded 90% by 4 weeks of culture. Whenever nonadherent cells were removed from the adherent cells and held with medium alone, they died within 48 h. Therefore, the lymphoid cells were to- tally dependent on adherent cells and/or their products.

2.5 Phenotype of adherent cells

2.5.1 Histochemical assays

Naphthyl acetate esterase and leukocyte peroxidase were detected with diagnostic kits from Sigma (90-A1 and 390-A). Acid phosphatase activity was shown with naphthol AS-BI phosphoric acid and fast garnet GBC salt (Sigma, 386-A), and the presence of alkaline phosphatase was demonstrated with

3.2 Comparison of suspended and adherent cell-associated lymphoid populations

Inspection of the cultures revealed that lymphoid cells were related to the adherent layer in different ways. Some were truly nonadherent, i .e. they were free floating in the medium or dispersed by gentle agitation or pipetting. Other, and often more abundant lymphocytes were very intimately associated with the adherent layer. (We refer to these as stromal-associ-

782 P. L. Witte, P. W. Kincade and V. VCtviEka Eur. J. Immunol. 1986.16: 779-787

ated cells.) They could be detached by vigorous mechanical agitation on a Vortex mixer or brief ( € 5 min) treatment with 0.02% EDTA.

A comparison of nonadherent and stromal-associated cells was made to investigate the possibility that the two populations differed in maturational or proliferative status. Cells were collected separately from three replicate dishes initiated with BALB/c marrow after 5 , 8 and 12 weeks of culture and analyzed with respect to total numbers, expression of a B-lymphocyte lineage marker (14.8) and synthesis andlor surface display of Ig chains (Table 1). Although numbers of nonadherent cells produced in individual cultures differed as much as fourfold, total numbers of lymphoid cells were generally comparable from dish to dish at a given time point (Table 1 and Fig. 1). Numbers of strornal-associated cells usually exceeded numbers of nonadherent cells but the two were similar in all other respects (Table 1).

LOW FALS measurements revealed that most lymphocytes harvested from the cultures were small, similar to small BM lymphocytes. However, substantial mitotic activity was indi- cated by DNA histograms and expression of the receptor for transferrin (Fig. 2). In several instances, cells of both types were subcultured onto fresh adherent layers and their subsequent growth found to be similar (unpublished observations). Therefore, we found no evidence that one population was less mature or more actively proliferative than the other.

Ceiiularity . x 14

.

L

! I l l 1 I I I I I I

2 4 6 8 10 1 2 0

100 90

- & 60 ? 50 240

Week of Culture

14 8+Cells

Table 1. A comparison of nonadherent (NAD) and stromal-associated (SAC) lymphoid cells from LTBMC-Ba)

Week No. cells/ % 14.8+ % sIgM' o/o pre-B cells

cultureNAD SAC NAD SAC NAD SAC NAD SAC of dish x B cells

5#1 2.0 9.1 47.1 23.0 10.6 3.9 11.0 13.3 #2 6.1 9.1 51.7 32.7 15.0 5.8 19.6 7.5 #3 3.7 8.1 48.2 44.7. 13.6 4.5 13.9 17.2

8 # 1 1.6 3.7 18.2 12.1 13.2 5.2 NDh' ND #2 1.0 4.5 21.6 25.1 6.2 3.7 ND ND #3 2.1 5.0 86.4 76.2 7.0 9.8 ND ND

12# l 3.3 4.7 2.8 0 .4 2.0 0 1.8 1.3 #2 5.0 6.1 38.6 31.8 18.4 9.0 57.9 31.2 #3 1.7 4.9 11.4 5.6 5.2 0 5.0 2.8

a) Replicate cultures from a single BM pool. b) ND = Not determined.

3.3 Interculture variability

We systematically conducted our studies with cells from separate cultures analyzed individually. This revealed two trends in the growth of recognizable B-lineage (14.8-bearing) lymphoid cells in LTBMC-B cultures. During the initial weeks of culture

100

60 50 40 :1 30

Pre - B Cells

I l l l l l l l l l l l

0 2 4 6 8 1 0 1 2

20 -I /

* I l l l l l l l l l l l

0 2 4 6 8 1 0 1 2

Week of Culture

s I g M + Cells

Figure 1. Total cell numbers and frequencies of B-lineage cell subsets in replicate cultures. Lym- phoid cells of BALB/c (closed circles. solid lines) or NZB/Bin (open circles, dashed lines) cul-

80 - intervals from 1 to 12 weeks. Cells from 3-5

D $ 6 0 - cultures were collected at each time point. The frequencies of particular

20 nonadherent lymphoid cells as this directly re- 10

0 2 4 6 8 10 12 0 2 4 6 8 10 12 that phenotype in the entire culture (see text). Bars represent mean number or frequency.

100 - . 90 - tures were harvested at

- 70 - 5 0 -

- phenotypes are shown for

-

- I T I T I I I I I I I flects the incidence of

Week of Culture Week of Culture


the B-lineage composite of each replicate culture tended to be similar (Fig. 1). Numbers of 14.8+, sIgMf and cp' cells declined rapidly, and at around 3 weeks of culture lymphoid cells of putative B lineage (14.8') began to grow rapidly in both frequency and number when individual culture dishes were examined. However, after 5 weeks of culture, the incidence of 14.8-expressing cells was highly variable and unpre- dictable. This variability persisted despite the consistency of total lymphoid cell numbers per culture and the overall "healthy" appearance of the cultures even at 12-14 weeks. The incidences of pre-B cells and B cells were less variable. The variable frequencies of 14.8+ cells were not attributed to loss of the common leukocyte antigen (T200) molecule as most lymphoid cells from either 14.8-high or 14.8-low cultures displayed a determinant common to this molecular family (detected with monoclonal Mlh39.18, data not shown).

In general, the variability of detectable early B-lineage cell frequencies under our culture conditions was highly reproduc- ible and occurred as early as 4 weeks of culture. This suggested that events which transpired during the initial weeks of

A. Figure 2. Comparison of the cell cycle status of nonadherent and stromal-associated lymphoid cells from individual LTBMC-B cultures. In (A) the FALS of the nonadherent and stromal-associated cells are compared with normal BM cells from a BALBic mouse 5 weeks of age. The cells shown (B) were stained with rat anti-mouse transferrin receptor mAb plus fluoresceinated mouse anti- rat Ig. Cells from the same cultures were analyzed for their content of DNA (C) . The vertical axis shows relative cell number and the hori- zontal axis, increasing light scatter or fluorescence inten- sity.

nonadherent

strornal- associated

r'i iz f i bone

0.

' '.. stromal-associated 16.6% L -control Cel 1

C. Go'Gi S G2.M

I ...: I

culture led to subsequent interculture variability in the nature of the lymphoid cells supported after approximately 4-6 weeks of culture.

3.4 Temporal stability of B-lineage cell populations

The variable expression of the 14.8 antigen (some cultures high, some low) after the initial rapid growth period implied that the frequency of 14.8' cells might change with time in individual cultures. Therefore, we serially examined the temporal stability of the B lineage profile within individual cultures. Data in the previous sections showed that the incidence of B-lineage cells present in the nonadherent cell fraction reli- ably predicted the frequency in the entire lymphoid cell population per culture. By sampling nonadherent lymphoid cells only, disruption to the culture was minimized and most stromal-associated cells remained undisturbed. In the follow- ing experiments, the fate of B-lineage cells was determined through sequential harvest of individual cultures.

Beginning at 6 weeks and then at biweekly intervals, nonadherent cells were collected from replicate cultures originat- ing from the same bone marrow pools (Fig. 3). As before, the incidence of 14At cells was highly variable. Moreover, gradual loss of 14.8 antigen expression was a general trend among lymphocytes acquired from most cultures. Pre-B cells and B cells tended to be in the minority (usually less than 20%) and again, their frequencies were less variable (Fig. 3). Thus, the most typical course for the long-term cultures was an apparent expansion of cells with an early phenotype (i.e. 14.8- cp-, sIgM-) while cells of a later phenotype (14.8+, cp-, sIgM-) diminished. This was not always the case and in one culture, the numbers of pre-B cells increased while total 14.8' cells declined (Fig. 3, #7) . A few of these cultures were main-

Figure 3. Temporal stability of B lineage cell subpopulations within individual cultures. All nonadherent cells were harvested biweekly with care to minimally disrupt stromal-associated lymphocytes. The number of lymphoid cells and frequencies of B lineage cells were

#3, --- #4, followed for 4 weeks. #1, -..-..- #2, . . . . . - # 5 , - a _ . - #6, ----- #7).

ularity Pre-B Cells 501 c

0 z ,! j ---nonadherent

I ! f j

, I

- -stroma!-associated

-bone marrow

S*G2* M = \ ~ ____. 23 7 %

......... ......_..... 25.9%

21.3%

.._._. --.-

_. . .. .P.

6 8 10 6 8 10

Week of Cul ture

115432.21 [I]

784 P. L. Witte, P. W. Kincade and V. VetviEka Eur. J. Immunol. 1986.16: 779-787

tained and evaluated again after an additional 4 weeks. The tendency to decreased 14.8- expression continued over this interval (data not shown). However, in other studies, subset compositions remained very stable when established cultures were put through several cycles of subculture [20]. These observations suggested that production of B-lineage cells in long-term cultures was not directly equivalent to what is thought to occur in vivo and an apparent dysregulation occurred during a discrete phase in the growth pattern.

before or after subculture (data not shown). These approaches made it possible to study well separated adherent cells with respect to phagocytosis, Fc receptor expression, morphology and histochemical profiles.

3.5 Characteristics of adherent nonlymphoid cells

Both normal and dysregulated B lymphogenesis from fresh BM and fetal liver tissues are affected by nonlymphoid cells that adhere to Sephadex G-10 and plastic [6, 71. Because the lymphoid cells of these LTBMC-B were dependent on a pleiomorphic adherent cell layer, one likely explanation for dish to dish variation in lymphoid subpopulations was corre- sponding differences in the supporting adherent stromal cells.

Two predominant nonlymphoid cell types were easily distinguished by morphology: numerous macrophage-like cells and undefined, large cells possessing highly spread cytoplasm and large, oval nuclei. Stromal-associated lymphoid cells were most often found clustered upon these very large cells (Fig. 4). The identities and differential numbers of adherent cells per dish were analyzed in two ways. First, cells m m v e d from the dishes were immediately cytocentrifuged onto slides, and see and, a portion of the cells were rep1ated Onto glass coverslips or glass chamber slides for 48 h. Morphologic and histochemical scores correlated closely whether cells were examined

Figure 4 , Jenner-Giemsa-stained field of a LTBMC-B at weeks grown on Thermanox coverslips (LUX). Two general adherent cell types are apparent: macrophage-like cells (M) and very large cells (LC) upon which the lymphocytes seem to cluster. Magnification x 170 (approx.).

Table 2. Characteristics of adherent cells from LTBMC-B

Culture”) % Esterase’ c/r Acid phosphatase +

(A) Macrophage-like adherent cellsd) 3 weeks #I 95.5 93.0 #2 94.4 96.7 #3 96.7 93.5 8 weeks #I 91.5 81.1 #2 95.1 89.5 #3 98.0 85.7

(B) Large. fibroblastic adherent cells 3 weeks # I 8.3 7.0 #2 25.0 5.6 #3 0 21.4 8 weeks #I 21.9 7. I #2 28.0 40.0 #3 26.3 11 .1

% Alkaline % Phagocytosis phosphatase’ HEMA EA(1gM)‘’ EA(lgG2,,)

particlesh’

2.9 77.2 33.6 71.3 3.5 71.3 40.4 76.0 5.6 81.5 30.3 73.5

4.3 59.9 42.5 60.9 3.2 58.8 48.1 68. I 0 62.9 42.6 66.7

58.8 None None None 66.7 None None None 60.0 None None None

55.0 None None None 63.6 None None None 66.7 None None None

rh Pinocytosis of EA(IgG21,) dextran ( 2 h)

30.7 97.0 31.5 91.2 59.5 94.3

30.4 96.8 28.0 96.7 32.9 95.7

None None None None None None

None None None None None None

a) Replicate cultures (#1-3) were harvested from two separate BM pools (3 weeks and 8 weeks). b) In two separate experiments not shown, > 90% of cells with macrophage morphology also phagocytosed latex beads. c) Ingestion of SRBC coated with mouse anti-SRBC mAb of IgM, IgG,, or IgG,, subclasses. The ingestion of SRBC alone was 13-30% of total

cells. Cells were positive if 3 or more RBC were phagocytosed. d) The percentages are based on total counts of cells of each morphology. At least 200 cells were scored per slide.


The majority of the adherent cells appeared to be typical BM macrophages (range 60-95%). These cells were highly phagocytic for latex and HEMA particles and strongly positive for nonspecific esterase and acid phosphatase but negative for peroxidase; most possessed Fc receptors (Table 2). Replicate cultures within individual experiments were not highly variable in their content of macrophages (Table 3).

As a group, the large cells had common characteristics which clearly distinguished them from macrophages. They were to- tally nonphagocytic of latex or HEMA particles or erythrocytes coated with Ig of various subclasses (Table 2). The two adherent cell subpopulations differed in pinocytic ability. While the macrophages readily pinocytosed fluorescent dextran, fluorescence was not observed in the large cells after 2 h of incubation. Between 50-70% of the large cells were alkaline phosphatase positive, whereas cells of macrophage- like morphology were not. Acid phosphatase and esterase staining was variable and weak with the large, nonphagocytic cells while it was strong and uniform with the macrophages. Again, a notable feature was the proportional similarity of large cells in cultures from a single group (Table 3). There- fore, the culture to culture variation in proportions of these two major types of adherent cells did not seem to correlate with the nature of the lymphoid cells they supported.

Table 3. Content of adherent cells in replicate cultures from different BM pools

Culture Total viable Macrophage-like Large, fibroblastic age“’ cellsldish cellddish cellsldish

x 10-6 x 10-6 x 10-o

1 week # I 1.8 1.6 (90.0%)h) 0.2 (10.0%) x2 1.8 1.6 (89.8%) 0.2 (10.2%) #3 1.2 ND ND

#1 1.6 I .4 (85.0%) 0.2 (15.0%) x2 1.2 I .o (84.9%) 0.2 (15.1%) x3 1 .s 1.2 (82.7%) 0.3 (17.3%)

3 weeks

8 weeks # I #2 #3

12 weeks #1 #2 #3 #4

I4t weeks X I #2 #3

0.9 1.4 2.0

1 .o 1.9 0.9 1.4

1.5 1 .5 1.5

0.5 (S9.5%) 0.9 (67.1%) 1.5 (77.5%)

0.9 (94.9%)

ND ND

1.8 (93.1%)

1.3 (88.5%,) 1.3 (90.1%) 1.4 (84.870)

0.4 (40.S%) 0.5 (32.9%) 0.5 (22.596)

0.2 (5.1%) 0.1 (6.9%)

ND ND

0.2 (1 1.5%)

0.2 (15.2%) 0.2 (9.9%)

164 weeks #1 3.4 2.5 (72.8%) 0.9 (27.2%) #2 1 .o 0.7 (71.0%) 0.3 (29.0%) #3 2.8 2.1 (74.7%) 0.7 (25.3%)

a) Replicate cultures (#1-4) were harvested from 6 separate experiments.

b) Percentage of total.

3.6 New Zealand Black (NZB) cultures

Studies from this laboratory have shown that NZB strain mice are markedly dysregulated with respect to B-lymphocyte precursors [7, 211. For this reason, we conducted parallel studies to those reported above using young NZB BM cells. Other than a tendency for lower lymphoid cell numbers, the results were generally similar to those obtained with 4-week-old BALB/c BM, including variability between cultures (Fig. 1).

4 Discussion

The findings reported here have several practical and theoreti- cal implications. (a) Successful growth of lymphocytes in LTBMC is more dependent on serum lot variations than previously emphasized. (b) It is inadvisable to make conclusions from LTBMC when pooled cells from separate flasks are studied. (c) Individual culture characteristics are principally determined during a discrete period of time. (d) The relative ratios of particular B-lymphocyte lineage cells in LTBMC are variable and substantially different from those in normal hemopoietic tissues. (e) In contrast to the extreme interculture variation observed in lymphocyte subpopulations, adherent stromal cell populations in replicate cultures appear remark- ably similar. (f) Early precursors, as well as their more mature progeny, are probably capable of initiating LTBMC and of extensive self renewal.

With a carefully selected lot of FCS, only cells of lymphoid morphology were present after 4 weeks in the nonadherent and stromal-associated fractions. These were comprised of a heterogeneous population including low frequencies of pre-B and B cells, which increased with time in some cultures, and proposed earlier B-cell progenitors, detectable with mAb 14.8 and which are the presumptive precursors of cy-containing pre-B cells [14]. Even with standardized conditions and replicate cultures from the same marrow suspension, variations in the compositions of individual dishes were striking. This was particularly so with respect to cells displaying an early phenotype. These were recognized by the 14.8 mAb but lacked synthesis of Ig.

The lymphoid cells in these cultures were clearly structured into two compartments: suspended lymphocytes and those firmly attached to the stromal layer. Vigorous mechanical dis- sociation or treatment with EDTA was needed to dislodge the latter lymphocytes. It is unclear in other reports which populations have been analyzed, although one study likened the interaction of stromal-associated lymphoid cells and adherent cells in LTBMC-B to the thymocyte-thymic nurse cell (TNC) association [21]. In contrast to TNC-associated thymocytes, which are thought to be specially sequestered arrangements of functional thymocytes [22], stromal-associated lymphoid cells did not differ in maturity or proliferative status from nonadherent lymphoid cells, and this intraculture dichotomy did not explain the interculture variability of 14.8 expression.

An analogous situation of variability may exist in Dexter cultures. Although cell numbers are relatively equal, interculture variation in the content of early hemopoietic progenitors has been described [23, 241, a phenomenon Crouse et al. term “flask homeostasis” [24]. The source of the variation is unknown but may be reflective of the renewal characteristics of the stem cells [25] or qualities of the microenvironment [23].

786 P. L. Witte, P. W. Kincade andV. VttviEka Eur. J. Immunol. 1986.16: 779-787

One hypothesis addressed in the present studies is that variability among the supporting adherent cells may affect the lymphocyte variability. We identified two adherent cell populations; macrophages were predominant and large, nonphagocytic cells were a less frequent but constant population. This conclusion is consistent with that of Dorshkind et al. [21] which was based on morphology. The two populations were functionally distinct (Table 2). The histochemical profile (primarily alkaline phosphatase content) and the morphology of the large, nonphagocytic cells suggests that they are related to marrow fibroblastic cells [26], possibly adventitial reticular cells [5, 261. Such cells appear to have a special capacity for the support of hemopoiesis in vitro [27]. Both adherent cell populations had striking similarity to stromal cells in Dexter- like cultures [27, 281. Some differences existed, such as the lack of fat cells in LTBMC-B [21], but the spatial arrangements of phagocytic macrophages juxtaposed to, and even beneath (V. V., personal observations), large fibroblastic cells were essentially identical to primary cultures of BM-adherent cells [28] and stromal cell lines [29] which support hemopoiesis. Which, if either, of the adherent cell types is necessary to maintain LTBMC-B is under investigation. We have observed that the adherent (i. e. fibroblastic)-lymphoid cell associations in NZB cultures were far less numerous than in BALBic cultures, concomitant with lower lymphocyte cellu- larity in the NZB cultures (Fig. 1).

Although we found no obvious correlation between the composite of adherent supporting cells and the variation in B- lineage lymphocyte numbers, we cannot eliminate the possibility that there may be important regulatory elements in BM that are present in limiting numbers. Their selective survival in individual cultures might differ greatly and the net result of their loss or expansion could have a major impact on the growth of particular B-lineage cell sets.

Many other explanations can be invoked to account for the interculture variability we observed. A major question that must be resolved is what cells actually establish themselves in culture. Does a population of only immature cells survive and give rise to all of the other cell types or do precursors at different stages of differentiation have self renewal potential ? Only a limited number of divisions are thought to normally occur as pre-B cells become B cells in BM [18]. If self renewal is controlled by extrinsic factors, then this regulation may break down in culture. Examples of several stages of B lineage cells have been successfully cultured, including lymphoid stem cells [30], pre-B cells [31] and splenic B-cells [32], under conditions which do not appear to support maturation. Noteably, low numbers of late stage B-lineage cells were always present in our cultures even at the time of cellular decline between 1-3 weeks. This suggests that lymphocytes of various developmental stages established but long-term self renewal was un- predictable.

Studies of the kind that have been performed do not ade- quately reveal population dynamics. The relative abundance of different sets of B-lineage precursors and B cells within hemopoietic tissues is predictable at different stages of normal development [33]. Mechanisms which normally control pre- cursor-product relationships in situ might not be operative in culture allowing cells with an “early” phenotype to mature at some unknown and variable rate in certain cultures. At least two examples of the dysregulation of B cell formation in marrow have been linked to excessive production of soluble prod-

ucts found in urine or serum [7, 81, and in NZB mice these aberrations may be mediated by an adherent regulatory cell [34]. Thus, the seeming dysregulation of the B cell composition in the long-term cultures of normal BM may present a unique situation analogous to some disease states.

It is also possible that selective outgrowth of particular cells types occurs. This would be consistent with reports that lymphoid cells propagated for prolonged periods tend to become pauciclonal [9]. Although lymphoid cells from LTBMC-B repopulate deficient animals without evidence of malignancy [31, 351, we and others [31] have noted that some cultures gained high numbers of pre-B cells lacking 14.8 expression (Fig. 3, culture #7). This phenotype is extremely rare in normal marrow but has been associated with human pre-B leukemic lines [36]. The question whether such LTBMC-B cells are products of a transformation event (or premalignant changes) or are normal cells “frozen” in differentiation should be resolved.

Many fundamental questions persist about the establishment and maintenance of lymphocytes in these cultures. Of primary importance would be the demonstration that the B-lineage developmental sequence is ongoing in long-term culture. Associations with adherent stromal cells are necessary to sus- tain the growth of lymphocytes but perhaps are insufficient for appropriate maturation in some cultures. Therefore, variability of culture phenotype may be experimentally manipulated to reveal distinct correlations with microenvironmental cells or products. Although this system remains complex, LTBMC-B should allow insight into both normal population dynamics and aspects of dysregulated B lymphocyte production as more elements of the culture system are defined.

The excellent technical assistance of Ms. Anna Henley is much appreciated. Ms. Margaret Robinson and Robbie Robinson provided very capable assistance with cell sorter analyses. We thank Dr. E. Ravache for sharing the D N A analysis protocol and Dr. Ken Landreth, Ms. Grace Lee and Michael Beare for helpful discussions during the course of this work. The assistance of Ms. Charlotte Matheson in the preparation of the manuscript is appreciated.

Received January 22, 1986.

5 References

1 Kincade, P. W., Adv. Immunol. 1981. 31: 177. 2 Osmond, D. G. and Batten, S. J., Am. J . Anat. 1984. 170: 349. 3 Trentin, J. J., in Gordon, A. S. (Ed.), Regulation of ffemato-

poiesis vol. 1, Appelton-Centur-Crofts, New York 1970, p. 161. 4 Weiss, L., Anat. Rec. 1976. 186: 161. 5 Westen, H. and Bainton, D. F., J. Exp. Med. 1979. 150: 919. 6 Kincade, P. W., Lee, G., Paige, C. J. and Scheid, M. P., J .

7 Jyonouchi, H., Kincade, P. W., Good, R. A. and Gershwin, M.

8 Landreth, K. S. , Engelhard, D., Beare, M. H., Kincade, P. W.,

9 Whitlock, C. A. and Witte, 0. N., Proc. Natl. Acad. Sci. USA

10 Dexter, T . M., Allen, T. D. and Lajtha, L. G., J. Cell. Physiol.

11 Dexter, T . M., J. Cell. Physiol. 1982. Suppl. 1: 87. 12 Whitlock, C. A., Robertson, D. and Witte, 0. N., J. Immunol.

13 Kincade, P. W., Lee, G., Watanabe, T., Sun, L. and Scheid, M.

Immunol. 1981. 127: 255.

E., J. Immunol. 1983. 131: 2219.

Kapoor, N. and Good, R. A., J. Immunol. 1985. 134: 2305.

1982. 79: 3608.

1977. 91: 335.

Methods 1984. 67: 353.

P., J. ImmunoE. 1981. 127: 2262.


14 Landreth, K. S. , Kincade, P. W., Lee, G. and Medlock, E. S. , J.

15 Kincade, P. W., Lee, G., Sun, L. and Watanabe, T., J. Immunol.

16 Lesley, J., Hyman, R., Schulte, R. and Trotter, J., Cell. Immunol.

17 Springer, T., Galfr2, G., Secher, D. and Milstein, C., Eur. J.

18 Landreth, K. S. , Rosse, C. and Clagett, J., J. Immunol. 1981.127:

19 Fornusek, C., Kopecek, J. and VgtviEka, V., Immunol. Lett. 1983.

20 Kincade, P. W., Witte, P. L. and Landreth, K. S. , Curr. Top.

21 Dorshkind, K., Schouest, L. and Fletcher, W. H., Cell Tissue Res.

22 Fink, P. J., Weissman, I. L., Kaplan, H. S. and Kyewski, B. A , , J.

23 Keller, G. M. and Phillips, R. A., Exp. Hematol. 1984. 12: 822. 24 Crouse, D. A., Mann, S. L. and Sharp, J. G., in Wright, D. G. and

Greenberger, J. S. (Eds.), Long-term Bone Marrow Culture, Alan R. Liss Inc., New York 1984, p. 211.

Immunol. 1983. 131: 572.

Methods 1981. 42: 17.

1984. 83: 14.

Immunol. 1978. 8: 539.

2027.

7: 29.

Microbiol. lmmunol. 1986, in press.

1985.239: 375.

Imrnunol. 1984. 132: 2266.

25 Mauch, P., Greenberger, J . S., Botnick, L., Hannon, E. and Hell-

26 Lichtman, M. L., Exp. Hematol. 1981. 9: 391. 27 Allen, T. D. and Dexter, T. M., Exp. Hematol. 1984. 12: 517. 28 Song, Z. X. and Quesenberry, P. J., Exp. Hematol. 1984.12: 523. 29 Zipori, D., Friedman, A., Tamir, M., Silverberg, D. and Malik,

30 Dorshkind, K. and Phillips, R. A., J . Immunol. 1983. 131: 2240. 31 Nagasawa, R., Kanagawa, O., Tittle, T. V. and Chiller, J. M., J.

32 Braun, J., J . Immunol. 1983.130: 2113. 33 Jyonouchi, H. and Kincade, P. W., J. Immunol. 1983.130: 1616. 34 Jyonouchi, H., Kincade, P. W., Landreth, K. S. , Lee, G., Good,

R. A. and Gershwin, M. E., J. Exp. Med. 1982. 155: 1665. 35 Kurland, J. I., Ziegler, S. F. and Witte, 0. N., Proc. Natl. Acad.

Sci. USA 1984. 81: 7554. 36 Landreth, K. S. and Kincade, P. W . , in Vitetta, E. and Fox, C. F.

(Eds.), B and T Cell Tumors: Biological and Clinical Aspects, UCLA Symposia on Molecular and Cellular Biology, Vol. XXIV, Academic Press, New York 1982, p. 57.

man, s., Proc. Natl. Acad. Sci. USA 1980. 77: 2927.

Z., J . Cell Physiol. 1984. 118: 143.

Immunol. 1985.135: 965.

Documents

Interculture variation and evolution of B lineage lymphocytes in long-term bone marrow culture