CROSS-REFERENCE TO GOVERNMENT GRANTS Research performed in support .of this application was supported in part by NIH grants CA03352 and CA42551 and performed at the Howard Hughes Medical Institute at the Leland Stanford, Jr. University. The United States government may have rights in any patent issuing on this application.

INTRODUCTION Technical Field

The field of this invention concerns hematopoietic cell differentiation in culture.

Background The thymus is the major site of differentiation of

T-lymphocytes. Bone marrow- or fetal liver-derived thymic progenitors migrate to the thymus where they undergo rapid proliferation and differentiation, a process known as thymic maturation. In the thymus, under the influence of the thymic stromal microenvironment, immature thymocytes may acquire various cell surface molecules including the MHC antigen co-receptors CD4 and CD8, and the T-cell receptor (TCR) alpha-beta or gamma-delta heterodi ers, associated with the invariant CD3 polypeptides. Based on their CD4 and CD8 expression, mature thymocytes are divided into four categories: CD4 ^" 8 ^" , CD4 ⁺ 8 ⁺ , CD4 ⁺ 8 ^~ and

CD4"8 ⁺ . The immature CD3"4"8" cells contain a subpopulation that proceeds through sequential CD3"4"8 ⁺ or CD3"4 ⁺ 8" large cells (a rapidly cycling population) and CD3 ^lo 4 ⁺ 8 ⁺ large thymocyte intermediates, before giving rise to the phenotypically and functionally mature CD3 ^hi 4 ⁺ 8" and CD3 ^hi 4 ^" 8 ⁺ T-cells, as well as CD3 ^hi 4 ⁺ 8 ⁺ small thymocytes that die in situ . During the process of thymic education, maturing T-cells acquire the property of responsiveness to foreign antigens presented in association with self-MHC molecules by the antigen presenting cells (APC) and of tolerance or non-responsiveness to self antigens (negative selection) . Interaction of CD3 ^l0 4 ⁺ 8 ⁺ cells with affinity for self-MHC complex with or without unknown self peptides may be required for self selection (positive selection) . Negative selection to Mis and VjS17-detected antigens occurs at a later stage when CD3 ^mcd 4 ⁺ 8 ⁺ cells are committed to either the CD4 (CD3 ^mcd 4 ⁺ 8'°) or the CD8 (CD3 ^med 4 ^lo 8 ⁺ ) lineage.

Although the lineage relationships between CD3"4"8" cells and the phenotypically and functioning mature T-cell subsets have been established, the precise nature of the cell interactions and the role of soluble factors involved in the process is poorly understood. One might predict that epithelial cells present within the thymic stroma play an important role in thymocyte development. For example, it has been proposed, but not directly shown, that one important function of these epithelial cells is the presentation of MHC proteins to maturing T-cells, which contributes to T-cell-repertoire (positive) selection.

It is therefore of substantial interest to be able to grow T-cells at various stages of differentiation, to investigate the various interactions associated with such maturation or selection, the factors involved in such maturation and selection, to evaluate cells as T-cell precursors, and to be able to grow T-cells, which may find use in therapy for the treatment of various diseases.

Relevant Literature

Thymic maturation is described by Adkins, et al . (1987) Ann . Rev. Immunol . 5, 323 and Fowlkes and Pardoll (1989) Adv. Immunol . 44, 207. A description of the thymus structure may be found in van Ewijk (1991) Ann . Rev.

Immunol . 9, 591, Butcher, eissman and Small (1984) Eur. J. Immunology 14, 936. Some of the cells of the thymus have been grown in tissue culture and partially characterized (Small, et al . [1989] Immunology 68, 371). Some epithelial cell cultures or epithelial cell lines have been shown to be effective in promoting in vitro growth and, at least, partial differentiation of either murine (Tatsumi, et al . [1990] Proc . Natl . Acad . Sci . USA 88, 642) or human (Denning, et al . [1989] J. Immunol . 142, 2985) thymic lymphocytes. CD3 ^" 4 ^" 8 ^~ thymocytes or putative prothymocytic murine T cell clones (Palacios, et al . [1989] EMBO J. 8, 4055). The heterodimeric cell surface antigen 1C11 which serves as a marker of a subset of CD3"4"8 ^" cells is described by Sen-Majumdar, et al . (1990) J. Immunol . 144, ill.

SUMMARY OF THE INVENTION Hematopoietic progenitor cells capable of differentiation and maturation in the T-cell lineage pathway are grown in co-culture. Thymic stromal cells are obtained by harvesting thymic tissue and removing thymocytes. Stromal cells are then grown in an appropriate nutrient medium in the presence of an unnatural stereoisomer amino acid. The medium is replaced without the unnatural amino acid and the cell population of interest added. The cell population may then be investigated after a predetermined period of growth, where particular cell populations can be obtained.

DESCRIPTION OF THE SPECIFIC EMBODIMENTS

In accordance with the subject invention, methods and compositions are provided for growing cells in

culture, where the cells are dedicated to the T-lymphocyte lineage. The methods permit the use of T-cell progenitor cells at various levels of maturation for differentiation into more mature cells, whereby the various agents and interactions involved in such maturation process may be studied, as well as the effect of agents, chemical, biological and physical, on such maturation process may also be studied. In this manner, one may produce T-cells at various levels of maturation, one can evaluate the effect of agents on the process of maturation, and one can isolate T-cell progenitors or T-cells at various stages of maturation. The method may also be used to investigate the ability of a mammalian cell population to induce cells in the T-cell lineage. The cells are grown in a culture medium under co-culture conditions, where thymic stromal cells are used for the co-culture, referred to as heterogeneous thymus stromal cells ("HTSC") . The thymic stromal cells can be obtained from a thymus of a particular species, at a relatively young age, or fetal, usually at an age less than about 25%, more usually less than about 15%, of the normal lifespan of the particular species. These cells will normally include various types of epithelial cells (cortical and medullary types, mesenchymal derivatives, dendritic cells, and macrophages) , the epithelial cells being of particular importance. The cells may be dispersed in appropriate medium by any convenient means, using physical separation, enzymatic separation, e . g. collagenase, trypsin/EDTA, etc . , or the like. The number of cells which will be used to seed a culture dish of 4 cm. size, will generally be the cells present in an about 4 wk. mature mouse.

The nutrient medium which is employed may be any conventional medium for growing mammalian cells, such as DMEM, RPMI, MEM, Iscove's media, etc . Included in the medium will be a fibroblast growth inhibitor, conveniently an unnatural stereoisomer amino acid, particularly an aliphatic amino acid, preferably a non-polar amino acid,

particularly D-valine. The D-valine will be present in from about 50 to 120 mg/L, usually about 90 mg/L.

The stromal cells are normally grown for at least about 2 weeks, more usually for at least about 3 weeks and not more than about 10 weeks, more usually not more than about 6 weeks, prior to use with the cells of interest. In some instances, it may be desirable to remove adherent cells from the primary culture in a fixed container, using trypsin and EDTA treatment and transfer them to another container, e . g. multiwell plates.

The stromal cells and the T cell progenitors may come from any convenient vertebrate, particularly mammalian, host, such as primate, including human, murine, ovine, porcine, equine, canine, feline, lagomorpha, bovine, avian, and the like, conveniently murine. The stromal cells which are employed need not be the same species as the cells of interest. Stromal cells from one species may therefore be used in combination with the same or different species, e.gr. murine stromal cells with human T-cell progenitors.

Prior to introducing the cell population of interest, the medium will normally be exchanged, so as to remove the presence of the unnatural stereoisomer amino acid. The cell population of interest may then be added to the stromal cell culture. The cell population may be conveniently suspended in an appropriate nutrient medium, such as Iscove's modified DMEM or RPMI-1640, normally supplemented with fetal calf serum (about 5-10%) , L-glutamine, a thiol, particularly 2-mercaptoethanol and antibiotics, e . g. penicillin and streptomycin. The cell population of interest may then be removed from the co-culture by any convenient means, e . g. gentle pipetting, at various times after addition. The cell population may include hematopoietic stem cells, bone marrow, CD34 ⁺ cells, CD34 ⁺ 38 ^" cells, CD34 ⁺ Thy-1 ⁺ cells substantially free of CD markers associated with dedicated (non-T-cell) lineages, particularly human thymocytes, peripheral blood

cells or fractions thereof, fetal liver cells, or the like. It is found that with T-cell progenitors, e . g. thymocytes, a substantial change in the cell population profile can occur within one day. The isolated population may be identified in a variety of ways. Flow cytometry may be employed, where the cells may be identified by employing fluorescent labeled monoclonal antibodies, specific for a particular marker. Panning may be employed for removing one or more subsets in the T-cell lineage. Affinity columns, magnetic beads, or other devices involving solid surfaces to which antibodies may be bound, may also be employed for isolating and/or identifying particular T-cell progenitor subsets. The subject co-cultures may be used in a wide variety of ways. The nutrient medium, which is a conditioned medium, may be isolated at various stages and the components analyzed. Separation can be achieved with HPLC, reversed phase-HPLC, gel electrophoresis, isoelectric focusing, dialysis, or other non-degradative techniques, which allow for separation by molecular weight, molecular volume, charge, combinations thereof, or the like. One or more of these techniques may be combined to enrich further for specific fractions. One can also provide for the isolation of a wide variety of subsets of T-cell populations, where the various subsets may have desirable functional characteristics associated with that level of maturation. T-cell subsets include CD3 ^x 4-8 ⁺ , CD3 ^x 4 ⁺ 8 ^" , CD3 ⁺ 4'8-, CD3 ⁺ 4 ⁺ 8 ⁺ , where x may be -, lo, med or hi, these markings in combination with TCR α, β or TCR δ , y, CD3"4 ^" 8 ^" 1C11 ⁺ , IL-2R or particular V/3, pgp-1, etc . The populations may be further divided in heat stable antigen (HSA) positive or negative, particularly HSA ⁺ 1C11 ⁺ , which includes a CD3" CD4" CD8 ^" population. The populations may be obtained in substantially pure form, greater than about 90 no.% or one

may use combinations thereof, combining subsets at the same or different levels of maturation.

The various subsets of cells can have a variety of applications. Particular subsets may be isolated as expanded populations and returned to a patient to provide T-cell competence. Antigen presenting cells may be used to activate particular T-cell subsets associated with a particular antigen. The T-cells may be autologous or allogeneic. In addition, the T-cells may be modified, where DNA may be introduced by various techniques, such as spheroplast fusion, electroporation, or the like, where the cells may then be used in a variety of ways, to study various features of T-cell development, use the T-cells in various therapies, such as the treatment of cancer, infection, or the like. Genes may be introduced for a variety of purposes, e . g . prevent HIV infection, provide recognition of a particular MHC antigen-peptide complex, suppress activation of a particular T-cell TCR, etc . Various techniques have been established for infusion of cells into a host, where the cells may be injected, introduced by catheter, infusion, etc .

The following examples are offered by way of illustration and not by way of limitation.

EXPERIMENTAL

Materials and Methods

Mice. 4-6-week-θld C57BL/Ka (Thy-1.2, Ly-5.1), C57BL/Ka BA (Thy-1.1, Ly-5.1), and C57BL6/K-Ly5.2 (Thy-1.2, Ly-5.2) mice were bred and maintained in an animal facility. mAbs and Fluorescent Reagents. The sources and fluorochrome modification of mAbs specific for CD4, CD8, Thy-1.1 and Thy-1.2, Ly5.1, Ly-5.2 and Pgp-1 are described in Spangrude, et al . (1988) Proc . Natl . Acad . Sci . USA 80, 5694. Sources and specificities of other mAbs were as follows: FITC-145-2C11 (anti-CD3, Boehringer Mannheim Corp., Indianapolis, IN), PE-conjugated GK1.5 (anti-CD4; Becton Dickinson and Co. , Mountain View, CA) , biotinylated

anti-TCR/α β (Pharmingen, San Diego, CA) . Avidin- conjugated Texas red (TR) , -FITC, -PE and allophycocyanin (AP) were purchased from Caltag Laboratories (South San Francisco, CA) . Rat hybridoma cell lines producing antibodies to

IL-2R (mAb7D4 [Malek, et al . (1983) Proc . Natl . Acad . Sci . USA 80, 5694]) and heat-stable antigen (HSA, mAb M169 [Springer, et al . (1978) Eur. J. Immunol . 8, 539]) were obtained from the ATCC (Bethesda, MD) . In all cell sorting experiments, mAb ICll (Sen-Majumdar, et al . [1990] J. Immunol . 144, 4111) was used as the hybridoma culture supernatant and its reactivity was monitored with TR-conjugated goat anti-rat Ig (Caltag Laboratories) . In some experiments, purified ICll, M-169 and IM781 (anti- Pgp-1) antibodies were biotinylated using

N-hydroxysuccinimide biotin (Sigma Chemical, St. Louis, MO) after standard conjugation procedures. The reactivity of biotinylated reagents was monitored with avidin- conjugated fluorochromes. HTSC (heterogeneous thymic stromal cell) culture system. Thymic stromal cells from 4-6-week-old C57BL/Aa (either Thy-1.1 or Thy-1.2) or C57BL/6J-Ly5.2 (Thy-1.2, Ly5.2) mice were cultured. Thymuses were removed aseptically and the majority of the lymphocytes dispersed by pressure through a stainless steel grid placed in

Dulbecco's Modified Eagles Medium (DMEM) . The remaining stroma was transferred as a single thymus to each plastic petri dish (3002 or 3001, Falcon Plastics, Oxnard, CA) and minced into small fragments. The fragments were spread over the dish and allowed to dry for a few moments to facilitate adherence. 4 ml Modified Medium were then added to 5 cm dishes (or 2.5 ml to 3 cm dishes) and usually replaced twice weekly. Cultures were maintained in a 37°C incubator with 10% C0 ₂ . These stromal cells were then grown in MEM containing D-valine, where the composition was prepared as follows. To each 100 ml was added 9.4 mg D-valine, 10 mg glutathione, 5 mg ascorbic

acid, 0.04 mg insulin, 58.4 mg glutamine, antibiotics (10 mg kana ycin and 2.5 mg gentamicin) at 10% heat- activated fetal calf serum (FCS) (Biological Industries) . In some cases, adherent cells from primary culture dishes were removed by trypsin-EDTA (Gibco Laboratories, Grand Island, NY) treatment and placed into 24-well plates (Costar, Cambridge, MA) . D-valine containing medium was withdrawn just before sorted thymocytes were added to the epithelial cell culture. Sorted thymocytes were resuspended in Iscove's modified DMEM or RPMI-1640 supplemented with 5-10% FCS, L-glutamine, 50 μM 2-ME and penicillin-streptomycin. Thymocytes were removed after 1 (18-20 h) or 2 (40-43 h) days of culture by gentle pipetting, were centrifuged and viability was determined before flow cytometric analysis.

Cell sorting and flow cytometric analysis. Freshly isolated thymocytes were stained for 2-, 3-, or 4-color analysis and the fluorescence was analyzed using a highly- modified dual laser FACS IV (Becton Dickinson & Co.) with 4 decads logarithmic amplifiers as described (Guidos, et al . [1989] Proc . Natl . Acad . Sci . USA 96, 7542 and Sen-Majumdar [1990], supra) . Dead cells were detected by adding 1 μg/ml propidium iodide (PI) , and gated out by setting an electronic gate to exclude Pi-positive cells. For 4-color analysis, dead cells were diminished by the usual scatter gating method, and the excitation wave¬ length of the dyelaser was raised from 590-650 nm. Contaminating stromal cells present in the thymocyte populations were eliminated by setting electronic gates on forward angle light scatter as well as on obtuse scatter, (granularity measuring index) . Fluorescent data were analyzed by using the FACS/DESK program and presented either in the form of a histogram or two-parameter probability plates (5%) . The sorting procedure for CD3 ^" 4 ^" 8 ^" and lCll ^hi CD3"4"8" thymocytes has been described previously (Guidos, et al . [1989] supra) . Briefly, normal thymocytes were stained

with biotinylated anti-CD4 and CD8, and thymocytes expressing CD4 and CD8 molecules were depleted by the use of avidin-conjugated paramagnetic beads (usually 10 μl of beads/5 x 10 ⁶ thymocytes) . The procedure was repeated, although in some experiments one round of bead separation was found to be sufficient. Finally, CD4 ^" 8 ^" thymocytes were stained with a mixture of avidin-phycoerythrin (AV-PE) (to detect cells expressing low levels of CD4 and CD8) and FITC-conjugated anti-CD3 antibody. PE- and FITC- positive cells were gated out in the cell sorter and

CD3"4"8 ^" cells were collected. To select the lCll ^hl subsets of CD3 ^~ 4"8 ^" cells, thymocytes were stained with mAb ICll followed by goat anti-rat Ig-conjugated TR. After blocking the remaining reactive sites of secondary antibody with normal rat serum, the thymocytes were treated as described above for total CD3"4"8" cell sorting. Electronic gates were set so that lCll ^hi cells (expressing high levels of TR) within the CD3"4"8" population were collected. A small aliquot of the sorted cells was reanalyzed to calculate the purity of the preparation and in all experiments described, the purity of sorted cell populations was greater than 98%.

Results Differentiation of CD3'4'8' thymocytes after 1 d culture on HTSC.

To analyze the ability of HTSC to promote thymocyte differentiation, CD3"4"8' thymocytes were isolated from normal adult C57BL/Ka or C57BL/Ka/Thy-l.l mice as described. After 1 day of culture, the viability of the input cells was 40-60% in the presence of HTSC, but only 20-25% of the cells survived in the medium alone. About 15-20% of the thymocytes recovered from the medium alone acquired low levels of CD8 antigen, but only a marginal percentage of cells were CD4 ⁺ 8 ⁺ . In contrast, in the presence of the stromal cells, about 7-14% (3 experiments) of the cells were CD4 ⁺ 8 ⁺ after 1 day of culture. The

thy ocytes recovered from the stromal layer had a lower percentage of CD4"8 ⁺ cells (6-9%) than those which were kept in medium or that were found after intrathymic injection in vivo (Guidos et al . [1989] supra) . Inspection of the 2-color probability plots revealed that the shift towards higher expression of CD8 reflects an overall increase in staining in the FITC (CD8) channel, rather than emergence of a bimodality of CD8 staining. The CD3"4"8" cells were cultured alone or on a monolayer of HTSC for 1 d. The stromal cell cultures used in the study had been in culture for at least 3-5 wk before use. This reduces the probability that there would be any thymocytes left. Nevertheless, to rule out the possibility of minor thymocyte contaminants, thymocyte donors were either Thy-1 oir Ly-5 congenic to the donors of the HTSC cultures. Almost all cells analyzed after 1 or 2 d of in vitro culture were found to be positive for the congenic marker of the input thymocytes.

Differentiation of CD3'4'8' thymocytes into mature cell subsets. CD3"4"8" thymocytes were cultured in HTSC for 2 d and phenotypic analysis was performed to assess the differentiation pattern in the presence of epithelial cells. Although the general pattern of differentiation in different experiments was similar, we observed that the percentages of different cell subsets varied from one experiment to another. The in vitro differentiation assay (1 d) showed that > 80% of the input cells differentiated into predominantly CD4 ^" 8 ⁺ (57.4%) cells, and lesser numbers of CD4 ⁺ 8 ⁺ (23.8%) and CD4 ⁺ 8 ^" (2.9%) thymocytes. CD3"4"8" thymocytes from Ly-5.1 donors were grown in HTSC of Ly-5.2 origin. The emerging differentiated thymocytes expressed the Ly-5.1 marker demonstrating their origin from the input thymocytes.

In a second experiment, we examined the percentage of differentiated thymocytes expressing the CD3 antigen after 2 d culture period.

As compared with day 1, when 3-6% cells were CD3 ⁺ , a significantly higher percentage (36%) of the cultured thymocytes on day 2 expressed CD3/TCR marker. The level of expression of CD3/TCR antigen receptor complex on thymocytes relates to their state of maturation in vivo . Based on the relative density of CD3, the in vitro differentiated thymocytes were subdivided into three categories: cells expressing little or no CD3 (CD3) , cells that are low to medium positive (CD3 ⁺ ) , and medium to highly positive (CD3 ²⁺ ) . CD4 vs. CD8 profiles of each of these three CD3 subsets were determined. About 25% of the CD3'4"8" cells were found to have differentiated in this experiment and acquired either CD4 or CD8 antigen. Most of the cells from the CD3 ^" category were CD4 ^" 8", and CD3 ⁺ cells included a significant subset that expressed both

CD4 and CD8 antigens. However, when the CD4/CD8 phenotype of the CD3 ²⁺ cells was examined, a significant number of CD4 ⁺ 8 ^" and CD4"8 ⁺ mature single-positive cells were found. In addition, CD4"8" and CD4 ⁺ 8 ⁺ subsets were also present in the CD3 ²⁺ population. These results suggest that the heterogeneous stromal cell culture in vitro is capable of inducing differentiation of immature CD3"4"8 ^~ thymocytes into mature subsets of T-cells normally present in the mouse thymus. While differentiation of immature thymocytes was consistently observed in the HTSC culture system, the percentage of mature thymocytes varied between experiments. In addition, primary or secondary cultures of HTSC contained various types of adherent cells, with significant variations among different culture dishes, not only in terms of presence or absence of the stromal cell type, but also between the proportions of various kinds of stromal cells. The dependency upon the survival and differentiation of CD3 ^" 4 ^~ 8 ^~ thymocytes on the variability of the stromal cell culture population is unknown. For example, as to viability, about 33% of input cells survived in experiment 1, while 21% of input cells

survived in experiment 2. Differences were also found as to the extent to which the input cells would differentiate in the in vitro HTSC culture system.

To ensure that the observed in vitro differentiation of thymocytes was not due to preferential proliferation of mature cell contaminants, the following experiment was performed. To the thymic stromal cell cultures (C57BL, Thy-l.l/Ly-5.1) was added cell sorter-purified CD3 ^" 4"8" (C57BL, Thy-1.2/Ly-5.1) plus 2% mature CD3 ²⁺ contaminants obtained from C57BL (Thy-1.2/Ly-5.2) thymocytes. In controlled cultures, no contaminants were added. 2 d later in vitro cultured thymocytes were analyzed. The data demonstrated that although > 80% of the cells were positive for Thy-1.2, only 4.2% of these cells expressed Ly-5.2 in the well where contaminants were deliberately added. Thus, it appears that the thymocyte maturation observed in the presence of HTSC was due to differentiation of immature precursor populations and not to preferential proliferation of mature contaminants present in the sorted cell preparation.

A distinct subset of CD3'4'8' thymocytes express high levels of ICll antigen. Howe and MacDonald ([1988] J. Immunol . 140, 1047) report heterogeneity within the CD4"8" double-negative thymocytes, with respect to their phenotypes and distinct potentials for differentiation. A mAb ICll that preferentially marks thymic progenitor cells including the CD3"4"8" subsets has been reported (Sen-Majumdar [1990], supra) . The 1C11 ⁺ CD3 ^" 4 ^" 8 ^" thymocytes constitute about 26% of the CD3 ^" 4"8 ^" population, and to determine whether these thymocytes coexpress other T-cell differentiation antigens, the expression of ICll with that of each stable antigen was determined by flow cytometry. Normal thymocytes were stained with a mixture of anti-CD4 ⁺ anti-CD8 (PE) , anti-CD3 (FITC) , ICll (AP) and anti-HSA (TR) . The fluorescent signal for CD4/CD8 and CD3 were gated out so that the counterplots represented ICll vs. heat stable antigen (HSA) of thymocytes, which were

negative for CD4, CD8 and CD3. Almost all 1C11 ⁺ CD3"4"8" cells exhibited high levels of HSA expression, although many additional HSA ⁺ cells failed to express ICll. Thus, a distinct subset of cells within the total CD3 ^" 4 ^" 8 ^' population is identified by the high expression of ICll and HSA.

Differentiation of 1C11 ⁺ CD3'4 ^~ 8" thymocytes to CD4 ^~ 8 ^~ and CD4 ⁺ 8 ⁺ cells after 1 d culture on HTSC. For 18-20 h, the cellular subset generated 8% of CD4 ^" 8 ⁺ and about 9% of CD4 ⁺ 8 ⁺ cells when cultured in the presence of HTSC. A significant percentage (8%) of cells expressed CD8 antigen when sorted cells were cultured in medium only. Whether this represents a background staining problem or true differentiation is unclear, although a small subset of CD3 ^" 4 ^" 8 ^" blast thymocytes had been reported to differentiate in vitro to CD4"8 ⁺ and CD4 ⁺ 8 ⁺ in the absence of thymic epithelial cells (Fawlkes and Pardoll [1989] Adv. Immunol . 44, 207). Occasionally, a small percentage (2-4%) of cells recovered from medium only on day 1 expressed both CD4 and CD8 antigens. A small percentage (3-4%) of thymocytes recovered from HTSC cultures on day 1 expressed low levels of CD4 antigen (CD4 ⁺ 8 ^" ) . Similar to the total of CD3 ^" 4 ^" 8 ^" thymocytes, 1C11 ⁺ cells in the presence of HTSC generated 7-10% of cells that were CD3 ^" and about 70% of those cells were CD4 ⁺ 8 ⁺ .

Generation of mature cell subsets from ICll ^"1" CD3'4"8' thymocytes. In the presence of thymic epithelial cells, the differentiation of 1C11 ⁺ thymocytes was further advanced when the culture period was extended to 2 d (40-43 h) . Although 20-32% of the 1C11 ⁺ CD3-4 ^" 8- input cells survived after 2 d of culture in HTSC, most (> 98%) of these cells were dead when cultured in the absence of the stromal cells. Phenotypic analysis of thymocytes recovered after 2 d of culture was performed as described previously. More than 50% of the surviving thymocytes were found to have differentiated to CD4-8 ⁺ (21%) , CD4 ⁺ 8 ⁺ (28%) and CD4 ⁺ 8" (4%) cells in the presence of the stroma.

Of the differentiated thymocytes, > 50% expressed varying levels of the CD3 antigen. Although CD4 ^" 8 ⁺ thymocytes were found to be distributed among CD3 ^" , CD3 ⁺ and CD3 ²⁺ subsets, most of the CD4 ^" 8 ^" were CD3 ²⁺ . The majority of the CD4 ⁺ 8 ⁺ cells belonged to the CD3 ²⁺ category, but only a small percentage of CD4 ⁺ 8 ⁺ thymocytes belonged to the CD3 ⁺ subset. The significance of such high numbers of CD4 ⁺ 8 ⁺ cells in the CD3 ²⁺ subset is not clear, but the existence of such cells in the normal thymus has been reported (Shortman et al . [1991] J. Exp. Med . 173, 323). The relative level of CD3 among the differentiated subsets obtained from 1C11 ⁺ CD3 ^" 4 ^" 8 ^" thymocytes was then determined. Although the CD4 ⁺ 8 ^" cells obtained after 2 d of culture expressed the highest levels of CD3, the CD4"8 ⁺ and CD4 ^" 8 ^" cells could be clearly divided into two categories: CD3 ⁺ and CD3 ²⁺ . All of these populations are present in the normal thymus, which suggests that thymic stromal cells in vitro can facilitate the differentiation of immature thymocytes into various types of phenotypically mature T-cells.

Immature blast cells within the thymus can be detected by their distinct scatter characteristics. About 50% of these blast cells express the ICll antigen. The cell size of ICll CD3 ^" 4 ^" 8 ^" thymocytes was compared after cell sorting and 1 and 2 d of culture on HTSC with that of total normal thymocytes. About 54% of the sorted cells were found to be blasts and the cell size gradually decreased when the thymocytes were cultured for 1 (34.9% blasts) and 2 d (19.6% blasts). These results along with the phenotypic profiles of the T-cells recovered after 2 d of culture in lCll ^hi CD3 ^~ 4 ^" 8 ^" cells on HTSC suggests that immature thymocytes are capable of responding to the differentiation-transmitting signals derived from the stromal cells in much the same way as observed with total CD3"4"8" thymocytes.

It is evident from the above results that one can investigate T-cell differentiation in culture and use the

culture medium for determining the effect of various agents, both chemical and physical, on viability, differentiation, formation and selection of different subsets, and the like. In addition, the medium can be used to identify various factors associated with the proliferation and differentiation of thymocytes. One can also employ cell populations from various sources or isolate specific subsets of T-cells to determine their potential for differentiation, response to agents, ability to response to antigen presenting cells and be activated toward particular targets, and the like.

All publications and patent applications cited in this specification are herein incorporated by reference as if each individual publication or patent application were specifically and individually indicated to be incorporated by reference.

Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be readily apparent to those of ordinary skill in the art in light of the teachings of this invention that certain changes and modifications may be made thereto without departing from the spirit or scope of the appended claims.