development of t cells in vitro from precursors in mouse bone marrow

CELLULAR IMMUNOLOGY 127,92- 104 (1990)

Development of T Cells in Vitro from Precursors in Mouse Bone Marrow

PATRICIA BENVENISTE, BRIAN S. CHADWICK, AND RICHARD G. MILLER

Ontario Cancer Institute and the Department ofImmunology, University of Toronto, Toronto, Ontario, Canada M4X IK9

Received September 28, 1989; accepted December 17, 1989

Bone marrow cells from 6- to 8-week-old athymic nude mice were depleted of nylon-wool adherent cells and cultured in vitro at low cell numbers (300 cells/well) in medium supplemented with a supematant from a thymoma cell line. About 1% of cultured cells grew. Pooled cultures contained cells expressing CD3 (52%), CD4 (37%), CDS (11%) Thy 1.2 (72%) MAC-l (43%) and J 1 Id (86%) but no cells expressing sIg. They also contained cells expressing mRNA for the LX,@, y, and 6 chains of the T cell receptor as assessed with C region probes using a sensitive dot blot assay. These cells appear to develop from progenitors which are CD3-. When pooled Day 10 cultures were depleted of nylon-wool adherent cells, the remaining cells were nearly all J 1 Id+, Thy 1.2*, MAC-l-, CD3+, and either CD4+CD8+; CD4+CDSK; CD4-CDS+, or CD4-CD8-; i.e., their surface marker patterns were reminiscent of those of thymocytes. We conclude that our culture system is enabling bone marrow precursors to commence differentiation down the T cell lineage in the absence of a thymic environment. 0 1990 Academic PISS, IW.

INTRODUCTION

Bone marrow contains myelolymphoid stem cells which can, within the bone marrow environment, produce progeny capable of seeding to the thymus and developing into T cells (l-2). Some of these stem cells may have become committed to the T cell lineage in the bone marrow (3-8). However, little is known about these cells. In particular it is not clear whether one single subpopulation or distinct subpopulations of bone marrow cells are capable of seeding the thymus (e.g., myelolymphoid stem cells and/or lymphoid committed cells) nor is it known whether the thymus environment is absolutely essential for all T cell precursors to develop into functional T cells. Thus, athymic nude mice have no thymus but can develop functional T cells, albeit at a slow rate (9-13). In normal mice, at least some T cells, particularly those restricted by class-I-MHC, may be able to develop via an extrathymic pathway (14-16).

We have sought culture conditions that support the development of T cells directly from bone marrow without the need for a thymic microenvironment. It was found that when bone marrow cells from either athymic nude ( 17) or normal ( 18) mice were depleted of nylon-wool adherent cells (NWAd)’ and then cultured at low cell number

’ Abbreviations used: FITC, fluorescein isothiocyanate; NWNAd, nylon-wool nonadherent; NWAd, nylon-wool adherent.

92

0008-8749190 $3.00 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

DEVELOPMENT OF T CELLS IN VITRO 93

( l-300 cells/well) in medium supplemented with supernatant from a thymoma cell line about 1% of cells grew. Growing wells contained myeloid cells, lymphoid cells, or both. Many cultures containing lymphoid cells also contained spontaneously cytolytic cells with specificity for targets bearing self-MHC ( 17, 18). Cultures from nude and normal mice produced comparable results whereas cultures from scid mice, known to have a defect at an early stage in lymphoid development, failed to develop either lymphoid cells or spontaneously cytolytic cells (18). The factor(s) in the thymoma supernatant responsible for growth is unidentified but neither IL-2 alone nor the supernatant from Concanavalin A-stimulated spleen cells is sufficient to support growth(17, 18).

In this study, we further characterize the cells in cultures grown from athymic nude bone marrow and their precursors for expression of T cell surface markers (CD3, CD4, and CDS) and for expression at the mRNA level of genes coding for the T cell receptor. We conclude that CD3+ cells are developing from CD3- precursors. The pattern of CD3, CD4, and CD8 expression has some resemblance to that seen in normal intrathymic T cell differentiation.

Mice

MATERIALS AND METHODS

C57BL/6-nu/nu (B6-nude) and littermate C57BL/6-nu/+ (B6) were bred in the animal facility of the Ontario Cancer Institute and used when 6-8 weeks old.

Cell Lines

Tumor cell lines RBL5 (T lymphoma), EL4 (T thymoma), P8 15 (mastocytoma) and Sp6/HL- 1 (IgM myeloma) were maintained by passage in vitro in a-MEM plus 10% FCS. The mouse epidermal dendritic cell line Dd was a kind gift of Dr. D. Ferrick (Ontario Cancer Institute, Toronto, Canada) and is Thy-l+, y/6+ (private communi- cation).

Culture System

This was as previously described (18). Briefly, B6-nude bone marrow cells were depleted of nylon-wool adherent cells by passage through a nylon-wool column (re- covery 5-10%) and cultured at 300 cells/well in 96-well round-bottom microtiter trays. Inclusion of a selected batch of ELCSN (supematant from an EL-4 thymoma cell line stimulated for 48 hr with 10 rig/ml of phorbyl myristic acetate) at final con- centration of 30% was critical. On Day 10, growing wells, typically containing 3 X 1 O3 to 3 X lo4 cells each, were scored by light microscopy and analyzed further either individually or after pooling.

Surface Marker Analysis

Hybridoma 145-2Cll (anti-CD3, Ref. (19)) was a gift of J. Bluestone (NIH, Bethesda, MD). Hybridomas GK 1.5 (anti-CD4, Ref. (20)) 53-6.72 (anti-CD8, Ref. (21)) Ml/70 (anti-MAC-l, Ref. (22)) 30-H12 (anti-Thy 1.2, Ref. (21)) and Jlld (anti-heat stable antigen, Refs. (23,24)) were obtained from ATCC (Rockville, MD). Hybridoma 145-2C 11 was grown as ascites in athymic nude mice, purified by ammo-

94 BENVENISTE, CHADWICK, AND MILLER

nium sulfate precipitation and directly conjugated with fluorescein isothiocyanate using standard procedures (25). In other instances, tissue culture supematants from the MAC- 1 and/or J 1 Id hybridomas were used directly with FITC-conjugated goat anti-rat Ig (Southern Biotechnology Assoc. Inc., Cat. No. 3010-02) or FITC conjugated goat anti rat IgM (Jackson Immuno Research Laboratories Cat. No. 112-015- 020) respectively, as second antibody. For CD3 labeling, mouse IgG (10 pg/ml, Sigma, Cat. No. l-538 1) was used to saturate Fc binding sites. Avidin-phycoerythrin (Becton-Dickinson Cat. No. 133 1) was used to label biotinylated antibodies (CD@. Labeled cells were analyzed on a Coulter Epics V flow cytometer using the 488-nm line from an argon laser.

mRNA Dot Blots

These were performed using a modified commercial dot blot apparatus and procedure exactly as previously described (26), the advantage of this procedure being that it is lo- to 1 OO-fold more sensitive than standard procedures. Briefly, cells ( lo5 or fewer) were transferred onto a very small area of a dry nylon membrane (Gene Screen Plus, New England Nuclear), lysed, fixed with formaldehyde, prehybridized for 24 hr and hybridized for 48 hr. Probes for (Y, ,f3, and y mRNA of the T cell receptor (gifts of T. Mak, Ontario Cancer Institute, Toronto, Canada), for immunoglobulin p mRNA (gift of G. Wu, Department of Immunology, U. of Toronto, Toronto, Canada) and for tubulin (gift of V. Ling, Ontario Cancer Institute) were all as described previously (26). The probe for the C6 chain of the T cell receptor was a 900 bp cDNA cloned into the EcoRI site of puC 12 (gift of T. Mak, Ontario Cancer Institute) and has been included in more recent experiments.

RESULTS

Expression of CD3, CD4, and CD8 on Bone Marrow Culture Cells

We first examined the ability of our culture system to generate cells with T cell surface markers from precursor cells in the bone marrow of B6 athymic nude mice 6-8 weeks old. Bone marrow cell suspensions were depleted of nylon wool adherent (NWAd) cells before culture. The remaining nylon wool nonadherent (NWNAd) cells (5- 10% of input) were cultured in microtiter trays (300 NWNAd cells/wells, 60 wells/tray, 20-40 trays per experiment) in the presence of a supematant derived from EL-4 thymoma cells. In agreement with previous results ( 17, 18) about 1 in 200 to 1 in 400 cultured cells grew and, after 10 days, nearly all wells contained 3000 to 30,000 cells, Growing wells were pooled and flow cytometry was used to look for the presence of cells with various cell surface markers.

Figure 1 (top set of 3 panels) shows fluorescence intensity histograms for CD3, CD4, and CD8 from one experiment. Included in the figure are results for B6-nu/+ thymocytes (positive control) and pooled culture cells labeled with FITC-conjugated goat anti-rat Ig (negative control). The pooled bone marrow culture cells appear to contain CD3+, CD4+, and CD8+ cells. In agreement with the results of others (29), the CD3 profile of the thymus control contains a large peak of relatively dim cells and a much smaller peak of fairly bright cells; in contrast, the bone marrow culture cells contain a single broad CD3 peak of intermediate modal intensity. The CD4 and CD8 profiles for bone marrow culture cells have much lower modal intensities than


CD3 CD4 CD8

tog,, fluorescence intensity

FIG. 1. Distribution of CD3, CD4, and CD8 in pooled Day 10 B6 nude bone marrow cultures (A, top three panels, solid line) and in B6-nude NWNAd bone marrow cells (B, bottom three panels, solid line). Curves measured at the same time for B6-nu/+ thymocytes (. . .) and for bone marrow culture cells stained with FITC-conjugated goat anti-rat Ig (---) are included in all panels. Each distribution is based on analysis of 10.000 viable cells and all curves are drawn to the same scale.

seen for thymus and contain a smaller proportion of cells, particularly that for CDS. Table 1 (first two lines) lists mean frequencies for cell marker-positive cells in Day 10 pooled cultures and thymus controls from nine independent experiments.

Expression of CD3, CD4, and CD8 on Progenitor Cells

On the basis of CD3, CD4, and CD8 expression, our cultures appear to contain T cells. Since at least half of the cells in the cultures bear one or more of these markers and since cell number increases lo- to loo-fold during culture (300 cells/well cultured, 3000 to 30,000 cells/well collected) some of these cells must have grown in the cultures. From what did they grow? Figure 1 (bottom 3 panels) presents fluorescence intensity profiles for CD3, CD4, and CD8 expression for the B6-nude NWNAd bone

TABLE 1

Cell Surface Marker DistributionY

Cell analyzed C57BLJ6 n CD3 CD4 CD8 MAC- 1 Thy. 1 Jlld

Day 10 pooled cultures nu/nu 9 52+ 10 31* 13 llf 5 43 * 3 72f 4 86 +1

Thymus nu/+ 9 61-t I 89kll 81+13 <I 95-c 4 94 _+l NWNAd nu/nu BM 11 5+ 1 <2 <2 50_+8 6? 2 91 +1 NWNAd lymph

node nu/+ 6 94t 7 56+ 3 37* I <0.5 97+ 1 0.6 + 2 NWNAd Day 10

pooled cultures nu/nu 4 61 f 13 13*26 68&28 <0.5 83 + 10 89 +l

’ Cells from C57BL/6, nu/nu or nu/+, were labeled with FITC-conjugated CD3, CD4, CD8, or Thy 1.2 antibodies or with supernatants from hybridoma cultures reactive against MAC- 1, or J 11 d, followed by FlTC-conjugated goat anti-rat IgG or IgM. Entries show mean + standard deviation of the percentages of cells in each subpopulation.


Thymus

,’ :: v,.: /,: I,:,’ ” ,..* ,i:::.:. ii; b/F

l!!!Ll

:.:::, .: . . . : : : : i’ ‘. : .i: ‘.. .A’ :::j:: ” .’

,.~,. :.’ : .:: .: . . :. ::j . . . :

1 10 100

NW NAd BM

CD3

Control

FIG. 2. Two-dimensional histograms of forward angle light scatter vs CD3 fluorescence intensity for B6- nude NWNAd bone marrow cells and B6 nu/+ thymocytes. Cells preincubated with FITC-conjugated goat anti-rat Ig were used as controls.

marrow cells put into culture. We were unable to detect cells expressing CD4 or CD8 (~2%) but we did detect a small subpopulation of dimly positive CD3+ cells. Figure 2 shows a two-dimensional forward angle light scatter versus CD3-fluorescence intensity analysis in which this population of CD3+ cells is clearly visible as a population of cells with forward angle light scatter and CD3 intensity corresponding approximately to that of CD3 dim thymocytes. Table 1 (line 3) shows pooled data from 11 experiments. The mean frequency of the CD3+ subpopulation is 5.1 + 1.2% and its mean fluorescence intensity is 8.3-fold lower than the bright peak of thymocytes.

We next tested directly whether the CD3+ cells in our starting B6-nude NWNAd bone marrow cell population gave rise to the cells bearing CD3, CD4, and/or CD8 in our cultures. B6-nude NWNAd bone marrow cells were sorted into CD3+ and CD3- subpopulations which were cultured as usual along with an unfractionated control. Multiple replicates of different cell numbers (10,30, 100,3OO/well) were also cultured so that growth frequency could be calculated using limiting dilution analysis as was done in previous studies ( 17, 18). Growing wells were pooled and analyzed for expression of CD3 and CD4. Table 2 shows results for one of four similar experiments. In this experiment, the CD3- and CD3+ sorted starting populations were, respectively, 99 and 85% pure on reanalysis. Both the growth frequency and CD3, CD4 marker expression were identical for CD3- and unfractionated populations. The growth frequency for CD3+ cells was about lo-fold lower and could be accounted for entirely by CD3- contaminants. These results suggest that the precursor cells giv-


TABLE 2

Comparison of Growth Frequencies and T Cell Markers on Pooled Day 10 Culture Cells Grown From CD3-Enriched or Depleted NWNAd Nude Bone Marrow Cells”

Cells cultured Growth frequency CD3 CD4

Unfractionated CD3- (99% purity)

CD3+ (85% purity)

l/362 56.6 39 l/323 62.6 31

l/3255 NT’ NT’

u Unfractionated, CD3- and CD3+ sorted fractions of NWNAd B6-nu/nu bone marrow cells were cultured with EL4 supernatant for 10 days. Growth frequencies were calculated for each fraction and growing cultures pooled for marker analysis as in Fig. I and line 1 of Table 2.

’ NT, not tested; not enough cells for analysis.

ing rise to CD3’, CD4+, and CD8 cells in our cultures are CD3- and, given flow cytometry limitations, would appear to also be CD4-, CD8.

Expression of CD3, CD4, and CD8 on NWNAd Bone Marrow Culture Cells

Not all the cells in our Day 10 cultures are T cells. Thus many cells have myeloid morphology and (Table 1, line 1) 43% of cells bear the marker MAC- 1, known to be primarily a myeloid marker (22). However, when pooled cells from the bone marrow cultures were depleted of NWAd cells, all MAC- I+ cells were eliminated while at the same time almost all CD8 cells were retained.

These NWNAd Day 10 cultured cells (7- 12% of input to the nylon-wool column) were further analyzed by flow cytometry. Figure 3 and Table I (line 5) show that mean frequencies of 60% or more were seen for CD3, CD4, and CD8 suggesting that many cells must carry two and at least some of all three of the markers. We therefore undertook two color fluorescence flow cytometry measurements for coexpression of CD4 and CD8 by these cells. As shown in Fig. 3 and Table 3, four subpopulations can be identified corresponding to double negative (CD4- CD8-), double positive (CD4+ CD8+) and both possible single positive (CD4+ CD8-; CD4- CD8+). The double-positive subpopulation was the dominant population both in culture cells and thymus control. The percentage of CD4+ CD8- and CD4- CD8+ single-positive cells was found to vary (Table 3). See Discussion. Approximately 60% of both NWNAd BM culture cells and thymocytes expressed CD3 in association with CD8 (data not shown). As an additional control, we analyzed NWNAd B6 nu/+ lymph node cells and did not detect any double-positive cells (Table 1, line 4).

Table 1 includes data for other cell markers which were measured in all or nearly all experiments. The frequency of Thy If cells was comparable to or greater than that of CD3+ cells throughout. Although looked for, sIg+ cells were not seen in our cultures (data not shown). The Jl Id marker is present on B cells, myeloid cells, and many thymocytes but absent on mature T cells (23, 24). We found it on nearly all bone marrow cells, bone marrow culture cells (before or after depletion of NWAd cells) and on thymocytes but not on NWNAd lymph node cells. The fact that Jl Id was present on 89% of NWNAd bone marrow culture cells, of which 60% express CD3, implies that a substantial fraction of the cells with T cell markers in these cultures must also carry J 11 d and distinguishes them from mature T cells.


D

1 1 10 loo

F E

loo loo CD4 CD4

10 10

1 1

1 10 100

CD8

1 10 loo

CD8

FIG. 3. Distribution ofCD3 (A), CD4 (B), and CD8 (C)in NWNAd pooled Day 10 bone marrow culture cells. Similar distributions for thymocytes (line with circles) are also presented. Bone marrow culture cells and thymocytes were labeled with FITC-conjugated CD3, CD4, CD8, and/or goat anti-rat Ig (control panel D). Dual fluorescence analysis of BM culture cells (E) and thymocytes (F) was performed using FITC- conjugated CD4 mAb and biotinilated CD8 mAb. Biotinilated antibody was developed using avidin-conjugated PE. Each analysis was performed on lo4 viable cells on an Epics Flow Cytometer. All curves are drawn to the same scale. Note that single-positive populations are largely accumulated in channels at or near the baseline using the Coulter Epics V System.

Expression of Genes Codingfor the T Cell Receptor

In order to detect the expression of genes coding for the T cell receptor in the rather small numbers of cells provided to us by our bone marrow culture system, we first developed a sensitive mRNA dot blot technique that allowed the detection of a, /3, y, and (more recently) 6 mRNA in as few as 3000 cells (26). Results of probing duplicate samples of 1 O5 cells from d 10 pooled bone marrow culture cells and from four control cell lines are shown in Fig. 4 (top 5 rows). The pooled Day 10 bone marrow culture cells appear to contain mRNA transcripts for (Y, /3, y, and 6 genes of the T cell receptor but no message for the immunoglobulin Ccl gene. Note that control cell lines RBLS thymoma and a CTL line, known to express an cz/p T cell receptor, are (Y positive and 6 negative; Sp6/HL- 1, an @I-producing hybridoma, is both (Y and 6 negative; whereas Dd, a y/S T cell receptor expressing dendritic epidermal cell line, is (Y negative


TABLE 3

Subpopulations in NWNAd Pooled Culture Cells and Thymus”

Day 10 NWNAd pooled cultures

Subpopulations Expt 1 Expt 2 Expt 3 Expt 4 Thymush

CD4-CD8 32 3.9 4.6 12 6k4 CD4+CD8+ 44 69 30 51 81?8 CD4+CD8- 13 22 64 1.5 lo* 3 CD4-CD8+ 10 3.6 1.0 30 3+1

a Nylon-wool nonadherent Day 10 pooled cultures and thymocytes were analyzed as in Fig. 3 for coexpression of CD4 and CD8. Entries show the percentages of cells in each subpopulation measured in nine independent experiments for the thymus (mean f standard deviation) and in four independent experiments for B6 nude-derived cultures.

b Entries are means of nine independent experiments.

and 6 positive. These results suggest that there may be both a//3 and r/6 T cell receptor-bearing cells in the d 10 pooled bone marrow cultures.

Pooled bone marrow culture cells were depleted of NWAd cells and stained with CDCFITC and CD%PE as in Fig. 3 and sorted into CD4+, CD8+ (d+); CD4-, CD8- (d-); and CD4+, CDK plus CD4-, CD8+ (s+, sorted together). These were then ana-

dl0 pooled colonies i

,’

RBL, ,.:

CTL i

d

SpG/HL-I

Dd

o! P v Tub

CY P Y P Tub

dl0 NW pooled colonies

d+d-s d+d-s d+d-s d+d-s d+d-s

FIG. 4. Expression of mRNA for T cell receptor genes in pooled NWNAd nude bone marrow culture cells. Duplicate samples of lo5 cells from Day 10 pooled cultures and control cell lines (top 5 rows) or single samples of 2 X lo4 cells from Day 10 cultures depleted of NWAd cells and sorted into d* (CD4+CD8+), d- (CD4-CD8-), or s+ (CD4+CD8- and CD4-CD8+) (bottom row) were probed for mRNA expression as described under Materials and Methods. Control cell lines were the RBLS T cell tumor (o~+@+y+Kc(-), a cloned CTL line (a+@+y+d-p-), the myeloma Sp6/HL-1 (a-fl-r’Kr+), and an epidermal dendritic cell line, Dd, (ol-~+-y+i?~Q.


lyzed for expression of a, /3, y, and p (Fig. 4, bottom row). The d+ and s+ populations expressed (Y, p, and y but no ;CL, with the d+ signals being weaker than the s+ signals; the d- population produced even weaker signals, with no evidence for an (Y signal. Thymocytes were sorted and analyzed in the same way in parallel and gave comparable results (not shown). In all of these experiments, we also probed for tubulin mRNA to give an indication of the total amount of mRNA present.

During the development of the dot blot procedure, we established that detection of a signal for LY, /I, or y genes of the T cell receptor required about 3 X lo3 cells from a T cell tumor. In the results of Fig. 4 each dot was produced from lo5 or 2 X lo4 cells. Since positive signals were seen, a substantial fraction of the cells must contain message for T cell receptor genes, a conclusion consistent with the percentages pre- dicted by flow cytometry. As single cultures contain 3 X 103-3 X IO4 cells, it might be possible to detect message directly in the potentially clonal populations, and pre- liminary experiments confirmed that we could detect (Y, /3, and y transcripts in single cultures. However, the signals obtained were so close to background that we have not pursued the matter.

Can message expression be detected in the bone marrow cells used to seed the cultures? Figure 5 shows results for NWNAd bone marrow from three nude mouse strains: B6, FI (B6 X RNC), and Balb/c. T cell receptor mRNA coding for Ca and C6 genes were detected in all three. Note that control cell line RBLS is LY+ 6- and control cell line Dd is (Y- 6+ as expected.

DISCUSSION

We have previously described an in vitro liquid culture system that supports the growth of nylon-wool nonadherent (NWNAd) nude ( 17) and normal ( 18) bone marrow cells. We documented the presence, in some of these cultures, of self-MHC-reactive cells and also showed that the culture conditions were permissive for the development of lymphoid T cells. We have here characterized both the precursors and the T cell progeny observed in these cultures.

We found that nude NWNAd bone marrow cells were about 5% CD3+ but that there were no detectable (~2%) CD4+ or CD8+ cells (Fig. 1, Table 1). The removal of CD3+ cells did not affect growth or T cell differentiation patterns observed in Day 10 cultures (Table 2). Further, these CD3+ cells showed little growth under our culture conditions (Table 2), suggesting that the precursor giving rise to T cells in our cultures is a CD3- cell.

Day 10 cultures derived from either unfractionated or CD3-depleted NWNAd nude bone marrow cells contained T cells on the basis of cell surface marker expression (CD3, CD4, CD8 Tables 1-3, Figs. l-3) and T cell receptor mRNA expression (Figs. 4 and 5). It is known that some cells can express truncated message for TcR genes and that TcR are not expressed on the surface of such cells. The whole cell dot blot technique used does not allow one to distinguish between truncated and functional mRNA transcripts. However, cell surface expression of CD3 appears to require coordinate expression of a complete TcR dimer, either a/P or y/S (27), so that at least some of the mRNA signal we detected is likely to code for a complete message. Since both CLY and C6 mRNA signals were present in nude Day 10 pooled cultures as well as in individual cultures, and since a large fraction of those cells express CD3 on the surface, it is likely that both TcR (Y/P and TcR y/S bearing cells have been generated


B6

F, (bx k)

Balb/C

RBL5 Dd

FIG. 5. Expression of T cell receptor gene mRNA in NWNAd B6 nu/nu, F, (B6 X RNC) nu/nu and Balb/c nu/nu bone marrow cells. Three or four replicates of 10’ bone marrow nylon-wool nonadherent cells were transferred onto a nylon membrane filter using a modified dot blot apparatus. Filters were hybridized with TcR Ca or C6 probes. Titrations of control cell lines RBLS (a+fl+y’b-) and epidermal dendritic cell lines Dd (a-fl+y+d+) were analyzed in parallel.

in Day 10 cultures. We conclude that some of the cells developing in the cultures are lymphoid T cells. On the basis of expression of MAC-l (22) myeloid cells are also developing in the cultures.

To further characterize the T cells in the cultures we depleted d10 cultures of MAC- l+ cells using their nylon-wool adherence property and characterized the remainder for expression of CD3, CD4, CD8, and J 1 Id. A subpopulation of cells with the follow- ing characteristics was obtained:

1. Cell surface expression of CD3, CD4, and CD8 antigens in a pattern reminiscent of thymocytes (28-3 1).

2. Expression at the mRNA level of the same TcR genes in the same subsets as seen in thymus (32-34).

3. Expression of the J 11 d antigen, seen on most thymocytes but not on mature T cells (23,24).

Although the above points emphasize the similarity to thymocytes, there are differences:


1. In thymus, CD3 is distributed bimodaly, with a large population of dim cells and a small population of bright cells. Cells in our cultures express low to intermediate levels; the high subset is absent.

2. Most CD4, CD8 double-positive thymocytes express a relatively high level of CD4 and CD8; almost no cells in our cultures achieve this level.

Taken together, the data might suggest that T cell differentiation is not proceeding to completion in our cultures. The results may also reflect in part unique features associated with the nude mutation.

All attempts to carry our cultures longer than 10 days have been unsuccessful. By Day 12, cell number and cell viability have always fallen significantly. This has hin- dered our attempts to get data on single clones rather than on pooled cultures as reported here.

Reconstitution of Day 14 fetal thymus organ cultures with Day 14 thymocytes has been successfully used and has been proven suitable for the generation of the four major thymocyte subsets defined by the expression of CD4 and CD8 markers. How- ever, variations were found within thymus organ cultures in the relative percentages of CD4+ and CD8+ single-positive cells (35). Thus, in some instances, an absence of the CD4+ CD8- subset was reported. We describe similar variation in the ratios of single-positive cells bearing either CD4 or CD8 markers (Table 3). Although double- positive cells were found in four of four experiments, few single-positive cells were detected expressing CD4+ CD8- in one experiment (Table 3, Expt 4) and few single CD8+ CD4- were observed in another (Table 3, Expt 3). Assuming that our in vitro culture system is supporting the differentiation of immature lymphoid precursor cells, minor variations in culture conditions could have major effects on cell-cell interactions and/or growth factor release leading to T cell differentiation.

Growth in our cultures is taking place in the absence of a thymic environment. However growth does depend critically on factor(s) secreted by the EL-4 thymoma ( 17, I8), i.e., factors from a cell of thymic origin. Whether it also depends upon the cellular environment provided by other cells also growing in the cultures remains to be determined. Recently Hurwitz et al. (36) reported an in vitro culture system that starts with normal bone marrow and that supports the differentiation of T cells in a manner also reminiscent of thymic differentiation. Normal bone marrow cells, depleted of Thy If, Ly I+, CD4+, and CD8+ cells were cultured with irradiated splenic feeder cells and a supematant derived from rat spleen cells incubated with Con A. T cell hybridomas established from these cultures at different times after initiation showed T cell receptor gene rearrangements similar to those seen in fetal thymocytes and/or adult thymocytes at corresponding time points. It is interesting to note that “thymic like” differentiation of bone marrow precursor cells is observed in both sys- tems in the absence of a thymic environment.

We found that about 5% of cells in NWNAd nude bone marrow were CD3+ (Fig. 2, Table 1). These cells, also observed in NWNAd normal bone marrow (data not shown), expressed approximately IO-fold less CD3 than bright thymocytes and had the peculiarity of being small in size as defined by light scatter. Since CD3 determi- nants are invariably associated with either an (Y/P or a y/S dimer (27), it is likely that this subset contains cells belonging to both these lineages, given the expression at the mRNA level of CLY and C6 transcripts in nude bone marrow (Fig. 5). Although the elimination of these CD3+ cells from NWNAd nude bone marrow did not affect the


differentiation patterns seen in the T cells grown under our culture conditions, we have found in another study (37) that they can be induced to grow with anti-CD3 mAb coupled to a solid matrix (but not by free soluble anti-CD3 mAb) in medium containing IL-2 and that they give rise to CD3+ CD4- CD8+ cells capable of lysing a broad range of target cells. The same culture conditions did not support the growth of CD3- cells.

It is not clear whether the CD3- progenitors in our cultures are CD4+ or CD8+. Although we did not detect CD4+ or CD8+ cells in nude bone marrow in this study, in the study mentioned above (37) which included an additional separation procedure, CD3- CD4+ and CD3- CD8+ cells were detected. It is interesting to note that spleen colony-forming units (CFU-S) from bone marrow carry low levels of CD4 (38). It is also interesting to note that CD3- CD8+ cells can be the precursors of CD4+ CD8+ in thymus (39) although we have shown (unpublished) that sorted CD3- CD8+ cells isolated from nude bone marrow do not grow under the culture conditions described here.

In conclusion, our in vitro culture conditions appear to support the growth and differentiation of T cell from nude bone marrow precursors. Some of the cells produced showed a pattern of differentiation reminiscent of thymic development in spite of the absence of a thymus environment. Whether their precursors are totipotent stem cells or cells restricted to the T lineage remains to be determined. Whether these precursors would normally seed to the thymus or are on an extrathymic differentiation pathway also remains to be determined. Since the culture conditions can be manipulated at a clonal level, the system offers the unique possibility of analyzing precursor-product relationships in developing clones of T cells.

ACKNOWLEDGMENTS We thank J. Sheldon for help with the flow cytometric analysis and sorting, T. W. Mak for the T cell

receptor probes, G. E. Wu and C. Paige for the immunoglobulin probe, and V. Ling for the tubulin probe. This work was supported by the National Cancer Institute of Canada. P.B. was a Terry Fox Fellow.

REFERENCES 1. Moore, M. A. S., and Owen, J. J. T., J. EXQ. Med. 126,7 15, 1967. 2. Wu, A. M., Till, J. E., Siminovitch, L., and McCulloch, E. A., J. Exp. Med. 127,455, 1968. 3. Abramson, S., Miller, R. G., and Phillips, R. A., J. Exp. Med. 145, 1567, 1977. 4. Dyer, M. J. S., and Hunt, S., J. Exp. Med. 154, 1164, 198 1. 5. Monissey, P. J., Kruisbeek, A. M., Sharrow, S. O., and Singer, A., Proc. Natl. Acad. Sci. USA 79,2003,

1982. 6. Dick, J. E., Magli, M. C., Huszar, D., Phillips, R. A., and Bernstein, A., Cell42,71, 1985. 7. Keller, G., Paige, C., Gilboa, E., and Wagner, E. G., Nature (London) 318, 149, 1985. 8. Lemischka, I. R., Raulet, D. H., and Mulligan, R. C., Cell 45,9 17, 1986. 9. Gillis, S., Union, N. A., Baker, P. E., and Smith, K. A., J. Exp. Med. 149, 1460, 1979.

10. Hunig, T., and Bevan, M. J., J. Exp. Med. 152,688, 1980. 11. Miller, R. G., Deny, H. R., and Sajeant, B. J., J. Immunol. 130,63, 1983. 12. MacDonald, H. R., Blanc, C., Lees, R. K., and Sordat, B., J. Immunol. 136,4337, 1986. 13. Maleckar, J. R., and Sherman, L. A., J. Immunol. 138,3873, 1987. 14. Lake, J. P., Andrew, M. E., Pierce, C. W., and Braciale, T. J., J. Exp. Med. 152, 1805, 1980. 15. Kruisbeek, A. M., Sharrow, S. O., and Singer, A., J. Immunol. 130,1027, 1983. 16. Kast, W. M., de Waal, L. P., and Melief, C. J. M., J. Exp. Med. 160, 1752, 1984. 17. Reimann, J., and Miller, R. G., J. Exp. Med. 158, 1672, 1983. 18. Benveniste, P., and Miller, R. G., J. Immunol. 136,4399, 1986. 19. Leo, O., Foo, M., Sachs, D. H., Samelson, L. E., and Bluestone, J. A., Proc. Natl. Acad. Sci. USA 184,

1374, 1987.


20. Dialynas, D. P., Wilde, D. B., Marrack, P., Pierres, A., Wall, K. A., Harran, W., Dtten, G., Loken, M. R., Pierres, M., Kappler, J., and Fitch, F. W., Immunol. Rev. 74,29, 1983.

21. Ledbetter, J. A., and Herzenberg, L. A., Immunol. Rev. 47,63, 1980. 22. Springer, T., Galfre, G., Secher, D. S., and Milstein, C., Eur. J. Immunol. 9,301, 1979. 23. Bruce, J., Symington, F. W., McKearn, T. J., and Sprent, J., J. Zmmunol. 127,2496, 198 1. 24. Crispe, N., and Bevan, M. J., J. Immunol. 138,2013, 1987. 25. Goding, J., Biochem. Immunol. pp. 59-103,2nd ed. Academic Press, Toronto, 1986. 26. Benveniste, P., Deny, H., Wu, G. E., Paige, C. J., and Miller, R. G., .I. Immunol. Methods 107, 165,

1988. 27. Clevers, H., Alarcon, B., Wileman, T., and Terhorst, C., Annu. Rev. Immunol. 16,629, 1988. 28. Bluestone, J. A., Pardoll, D. M., Sharron, S. O., and Fowlkes, B. J., Nature (London) 326,87, 1987. 29. Scollay, R., Bartlett, P., and Shortman, K., Immunol. Rev. 82,79, 1984. 30. Scollay, R., Smith, J., and Stauffer, V., Immunol. Rev. 91, 129, 1986. 31. Roehm, N., Herron, L., Cambier, J., Diguisto, D., Haskins, K., Kappler, J., and Mat-rack, P., Cell38,

577, 1984. 32. Snodgrass, H. R., Kisielow, P., Kiefer, M., Steinmetz, M., and von Boehmer, H., Nature (London)

313,592,1985. 33. Collins, M. K. L., Tanigawa, G., Kissonerglis, A. M., Ritter, M., Price, K. M., Tonegawa, S., and

Owen, M. T., Prof. Natl. Acad. Sci. USA 82,4503, 1985. 34. Pardoll, D. M., Fowlkes, B. J., Lechler, R. I., Germain, R. N., and Schwartz, R. N., J. Exp. Med. 165,

1624, 1987. 35. Kingston, R., Jenkinson E. J., andOwen, J. T., Nature(London) 317,81 l-813, 1985. 36. Hurwitz, J. L., Samaridis, J., and Pelkonen, J., Cell52,821, 1988. 37. Benveniste, P., Chadwick, B. S., Miller, R. G., and Reimann, J., J. Zmmunol. 144,411, 1990. 38. Fredrickson, G. G., and Basch, R. S., J. Exp. Med. 169, 1473, 1989. 39. Nikolic-Zugic, J., Moore, M. W., and Bevan, M. J., Eur. J. Zmmunol. 19,649, 1989.

development of t cells in vitro from precursors in mouse bone marrow

Documents