WO1999050637A2 - Isolated multimeric complexes useful in analysis of t cells, peptides useful in making the complexes, and uses thereof - Google Patents

Abstract

The invention involves multicomponent complexes, which use useful in analysis of T cells. The complexes contain at least first and second binding partners, which bind to each other. The second binding partner engages a plurality of immune complexes, which comprise an NHC molecule, a β2 microglobulin molecule, and a peptide. Preferably, there are at least four of these immune complexes per multicomponent complex. These can be used to determine or to isolate cytolytic T cells.

Description

ISOLATED MULTIMERIC COMPLEXES USEFUL IN ANALYSIS OF T CELLS, PEPTIDES USEFUL IN MAKING THE COMPLEXES, AND USES THEREOF

RELATED APPLICATION

This application is a continuation in part of Serial No. 09/049,850, filed on March 27,

1998, and incorporated by reference in its entirety.

FIELD OF THE INVENTION

This invention relates to multicomponent complexes which are useful in analysis of T

cell populations.

RELATED ART

Certain aspects of this invention will be found in published materials involving some of

the inventors. See Dunbar et al., "Characterization of Human Cytotoxic T Lymphocyte (CTL)

responses To Tumor Antigens Using Fluorogenic MHC Class I/Peptide Complexes", Tumor

Immunology 92 Suppl. Dec. 1997, reporting abstract 3.3 of the 5th Annual Congress of the BSI.

Also see Dunbar et al., "Direct isolation, phenotyping and cloning of law- frequency antigen

specific cytotoxic T lymphocytes from peripheral blood", Curr. Biol. (in press), but possibly

available via Internet on March 16, 1998. BACKGROUND AND PRIOR ART

The study of the recognition or lack of recognition of cancer cells by a host organism has

proceeded in many different directions. Understanding of the field presumes some

understanding of both basic irnmunology and oncology.

Early research on mouse tumors revealed that these displayed molecules which led to

rejection of tumor cells when transplanted into syngeneic animals. These molecules are

"recognized" by T-cells in the recipient animal, and provoke a cytolytic T-cell response with

lysis of the transplanted cells. This evidence was first obtained with tumors induced in vitro by

chemical carcinogens, such as methylcholanthrene. The antigens expressed by the tumors and

which elicited the T-cell response were found to be different for each tumor. See Prehn, et al.,

J. Natl. Cane. Inst. 18: 769-778 (1957); Klein et al., Cancer Res. 20: 1561-1572 (1960); Gross,

Cancer Res. 3: 326-333 (1943), Basombrio, Cancer Res. 30: 2458-2462 (1970) for general

teachings on inducing tumors with chemical carcinogens and differences in cell surface antigens.

This class of antigens has come to be known as "tumor specific transplantation antigens" or

"TSTAs". Following the observation of the presentation of such antigens when induced by

chemical carcinogens, similar results were obtained when tumors were induced in vitro via

ultraviolet radiation. See Kripke, J. Natl. Cane. Inst. 53: 333-1336 (1974).

While T-cell mediated immune responses were observed for the types of tumor described

supra, spontaneous tumors were thought to be generally non-immunogenic. These were

therefore believed not to present antigens which provoked a response to the tumor in the tumor

carrying subject. See Hewitt, et al., Brit. J. Cancer 33: 241-259 (1976).

The family of turn^" antigen presenting cell lines are immunogenic variants obtained by

mutagenesis of mouse tumor cells or cell lines, as described by Boon et al., J. Exp. Med. 152: 1184-1193 (1980), the disclosure of which is incorporated by reference. To elaborate, turn

antigens are obtained by mutating tumor cells which do not generate an immune response in

syngeneic mice and will form tumors (i.e., "tum⁺" cells). When these turn" cells are

mutagenized, they are rejected by syngeneic mice, and fail to form tumors (thus "turn^""). See

Boon et al., Proc. Natl. Acad. Sci. USA 74: 272 (1977), the disclosure of which is incorporated

by reference. Many tumor types have been shown to exhibit this phenomenon. See, e.g., Frost

et al, Cancer Res. 43: 125 (1983).

It appears that turn^" variants fail to form progressive tumors because they initiate an

immune rejection process. The evidence in favor of this hypothesis includes the ability of "turn "

variants of tumors, i.e., those which do not normally form tumors, to do so in mice with immune

systems suppressed by sublethal irradiation, Van Pel et al., Proc. Natl. Acad. Sci. USA 76: 5282-

5285 (1979); and the observation that intraperitoneally injected turn^" cells of mastocytoma P815

multiply exponentially for 12-15 days, and then are eliminated in only a few days in the midst

of an influx of lymphocytes and macrophages (Uyttenhove et al., J. Exp. Med. 152: 1175-1183

(1980)). Further evidence includes the observation that mice acquire an immune memory which

permits them to resist subsequent challenge to the same turn^" variant, even when

immunosuppressive amounts of radiation are administered with the following challenge of cells

(Boon et al, Proc. Natl, Acad. Sci. USA 74: 272-275 (1977); Van Pel et al., supra: Uyttenhove

et al., supra). Later research found that when spontaneous tumors were subjected to

mutagenesis, immunogenic variants were produced which did generate a response. Indeed, these

variants were able to elicit an immune protective response against the original tumor. See Van

Pel et al., J. Exp. Med. 157: 1992-2001 (1983). Thus, it has been shown that it is possible to

elicit presentation of a so-called "tumor rejection antigen" in a tumor which is a target for a syngeneic rejection response. Similar results have been obtained when foreign genes have been

transfected into spontaneous tumors. SeeFearonetal., Cancer Res.48: 2975-1980(1988) in this

regard.

A class of antigens has been recognized which are presented on the surface of tumor cells

and are recognized by cytolytic T cells, leading to lysis. This class of antigens will be referred

to as "tumor rejection antigens" or "TRAs" hereafter. TRAs may or may not elicit antibody

responses. The extent to which these antigens have been studied, has been via cytolytic T cell

characterization studies, in vitro i.e., the study of the identification of the antigen by a particular

cytolytic T cell ("CTL" hereafter) subset. The subset proliferates upon recognition of the

presented tumor rej ection antigen, and the cells presenting the tumor rej ection antigens are lysed.

Characterization studies have identified CTL clones which specifically lyse cells expressing the

tumor rejection antigens. Examples of this work may be found in Levy et al., Adv. Cancer Res.

24: 1-59 (1977); Boon et al, J. Exp. Med. 152: 1184-1193 (1980); Brunner et al, J. Immunol.

124: 1627-1634 (1980); Maryanski et al., Eur. J. Immunol. 124: 1627-1634 (1980); Maryanski

et al., Eur. J. Immunol. 12: 406-412 (1982); Palladino et al., Cane. Res. 47: 5074-5079 (1987).

This type of analysis is required for other types of antigens recognized by CTLs, including minor

histocompatibility antigens, the male specific H-Y antigens, and the class of antigens referred

to as "turn " antigens, and discussed herein.

A tumor exemplary of the subject matter described supra is known as P815. See DePlaen

et al., Proc. Natl. Acad. Sci. USA 85: 2274-2278 (1988); Szikora et al., EMBO J 9: 1041-1050

(1990), and Sibille et al, J. Exp. Med. 172: 35-45 (1990), the disclosures of which are

incorporated by reference. The P815 tumor is a mastocytoma, induced in a DBA/2 mouse with

methylcholanthrene and cultured as both an in vitro tumor and a cell line. The P815 line has generated many turn^" variants following mutagenesis, including variants referred to as P91A

(DePlaen, supra). 35B (Szikora, supra), and PI 98 (Sibille, supra). In contrast to tumor rejection

antigens - and this is a key distinction - the turn^" antigens are only present after the tumor cells

are mutagenized. Tumor rejection antigens are present on cells of a given tumor without

mutagenesis. Hence, with reference to the literature, a cell line can be tuπT, such as the line

referred to as "PI", and can be provoked to produce turn^" variants. Since the turn^" phenotype

differs from that of the parent cell line, one expects a difference in the DNA of turn^" cell lines

as compared to their turn" parental lines, and this difference can be exploited to locate the gene

of interest in turn^" cells. As a result, it was found that genes of turn^" variants such as P91 A, 35B

and PI 98 differ from their normal alleles by point mutations in the coding regions of the gene.

See Szikora and Sibille. supra, and Lurquin et al., Cell 58: 293-303 (1989). This has proved not

to be the case with the TRAs of this invention. These papers also demonstrated that peptides

derived from the turn^" antigen are presented by H-2^d Class I molecules for recognition by CTLs.

P91 A is presented by L^d, P35 by D^d and PI 98 by K^d.

PCT application PCT/US92/04354, filed on May 22, 1992 assigned to the same assignee

as the subject application, teaches a family of human tumor rejection antigen precursor coding

genes, referred to as the MAGE family. Several of these genes are also discussed in van der

Bruggen et al, Science 254: 1643 (1991). It is now clear that the various genes of the MAGE

family are expressed in tumor cells, and can serve as markers for the diagnosis of such tumors,

as well as for other purposes discussed therein. See also Traversari et al., Immunogenetics 35 :

145 (1992); van der Bruggen et al., Science 254: 1643 (1991) and De Plaen, et al.,

Immunogenetics 40: 360 (1994). The mechanism by which a protein is processed and presented

on a cell surface has now been fairly well documented. A cursory review of the development of the field may be found in Barinaga, "Getting Some 'Backbone': How MHC Binds Peptides",

Science 257: 880 (1992); also, see Fremont et al, Science 257: 919 (1992); Matsumura et al.,

Science 257: 927 (1992); Engelhard, Ann. Rev. Immunol 12:181-207 (1994); Madden, et al.,

Cell 75:693-708 (1993); Ramensee, etal, Ann. Rev. Immunol 11:213-244(1993); German, Cell

76: 287-299 ( 1994). These papers generally point to a requirement that the peptide which binds

to an MHC/HLA molecule be nine amino acids long (a "nonapeptide"), and to the importance

of the second and ninth residues of the nonapeptide. For H-2K\ the anchor residues are

positions 5 and 8 of an octamer, for H-2D^b, they are positions 5 and 9 of a nonapeptide while the

anchor residues for HLA-A1 are positions 3 and 9 of a nonamer. Generally, for HLA molecules,

positions 2 and 9 are anchors.

Studies on the MAGE family of genes and many other TRAPs have now revealed that

these are processed to peptides which complex with HLA molecules, leading to lysis of the cells

presenting these by cytolytic T cells ("CTLs").

Research presented in, e.g., U.S. Patent No. 5,405,940 filed August 31 , 1992, and in U.S.

Patent No. 5,571,711, as well as many other issued U.S. patents found that when comparing

homologous regions of various MAGE genes to the region of the MAGE-1 gene coding for the

relevant nonapeptide, there is a great deal of homology, and identification of a family of

nonapeptides, all of which have the same N-terminal and C-terminal amino acids. These

nonapeptides are useful for various purposes including their use as immunogens, either alone or

coupled to carrier peptides. Nonapeptides are of sufficient size to constitute an antigenic epitope,

and the antibodies generated thereto were described as being useful for identifying the

nonapeptide, either as it exists alone, or as part of a larger polypeptide. The preceding survey of the relevant literature shows that various peptides, usually eight,

nine, or ten amino acids in length, complex with MHC molecules and present targets for

recognition by cytolytic T cells. A great deal of study has been carried out on melanoma, and

melanoma antigens which are recognized by cytolytic T cells are now divided into three broad

categories. The first, which includes many of the antigens discussed, supra, (e.g., MAGE), are

expressed in some melanomas, as well as other tumor types, and normal testis and placenta. The

antigens are the expression product of normal genes which are usually silent in normal tissues.

A second family of melanoma antigens includes antigens which are derived from mutant

forms of normal proteins. Examples of this family are MUM-1 (Coulie, et al., Proc. Natl. Acad.

Sci. USA92:7976-7980(1955)); CDK4(Wόlfel, etal., Science269:1281-1284(1955));Bcatenin

(Robbins, et al., J. Exp. Med. 183: 1185-1192 (1996)); and HLA- A2 (Brandel, et al., J. Exp. Med.

183:2501-2508 (1996)). A third category, also discussed, supra, includes the differentiation

antigens which are expressed by both melanoma and melanocytes. Exemplary are tyrosinase

gplOO, gp75, and Melan A Mart-1. See U.S. Patent No. 5,620,886 incorporated by reference,

with respect to Melan-A. See Wδlfel, et al., Eur. J. Immunol. 24: 759 (1994) and Brichard, et

al., Eur. J. Immunol. 26: 224 (1996) for tyrosinase; Kang, et al, J. Immunol. 155: 1343 (1995);

Cox, etal, Science 264: 716 (1994); Kawakami, etal., J. Immunol. 154: 3961 (1995) for gp 100;

Wang, et al, J. Exp. Med. 183: 1131 (1996) for gp 75.

Cytolytic T cells ("CTLs" hereafter) have been identified in peripheral blood

lymphocytes, and tumor infiltrating lymphocytes, of melanoma patients such as those who are

HLA-A^*0201 positive. See Kawakami, et al, Proc. Natl. Acad. Sci. USA 91 :3515 (1994);

Coulie, et al, J. Exp. Med. 180:35 (1994). When ten HLA-A'0201 restricted Melan-A specific

CTLs were derived from different patients were tested, nine of them were found to recognize and react with the peptide Ala Ala Gly He Gly He Leu Thr Val, (SEQ ID NO:l), which consists of

amino acids 27-35 of Melan-A. (Kawakami, et al, J. Exp. Med 180:347-352 (1994)). Rivoltini,

et al, J. Immunol 154:2257 (1995), showed that Melan-A specific CTLs could be induced by

stimulating PBLs from HLA- A^* 0201 positive normal donors, and melanoma patients, using SEQ

ID NO: 1. The strength of this response has led to SEQ ID NO: 1 being proposed as a target for

vaccine development. Further research has shown that a decapeptide, i.e.,

Glu Ala Ala Gly He Gly He Leu Thr Val

(SEQ ID NO:2), is actually a better target than SEQ ID NO: 1. See U.S. Patent Application

Serial No. 08/880,963, filed June 23 1997 and incorporated by references.

One difficulty in the area of cancer immunology is a lack of reliable protocols which can

be used to identify and to quantify in vivo cytolytic T lymphocyte responses. As a result, it is

difficult to characterize immune response, and to monitor vaccine trials. The invention described

hereafter was developed to address these, and other issues in the field. It has been found that

analysis of cytolytic T cells is greatly facilitated by the use of complexes containing a plurality

of T cell targets. More specifically, the invention relies on the known avidity of two binding

partners, such as avidin or streptavidin and biotin for each other. It is well known that every

molecule of avidin/streptavidin can bind to four biotin molecules. The invention involves

constructs where the avidin/streptavidin-biotin system is used to form complexes containing

multiple targets for cytolytic T cells, i.e., a plurality of immune complexes which comprise an

MHC molecule, such as an HLA molecule, a B2 microglobulin, and a peptide which binds to the

HLA molecule. The complex is labelled, and can be used to isolate, or to determine, cytolytic

T cells of interest in a sample. How these and other features of the invention are achieved will be seen in the disclosure

which follows.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1 shows a correlation between frequency of cells positive for particular peptide containing

tetramers and cytotoxicity.

Figure 2 shows that CD8⁺ cells, isolated from bulk cultures and stimulated and expanded without

antigen had high tumoricidal activity.

Figure 3 presents data on determination of relative activity of peptides.

Figure 4 sets forth data showing that complexes of HLA-A2 and SEQ ID NO: 4 are more stable

than complexes of HLA-A2 and SEQ ID NO: 1.

Figure 5 presents data contributing to analysis of both polyclonal and monoclonal T cell

populations specific for tetramers of HLA- A^* 0201 and SEQ ID NO: 4.

Figure 6 presents phenotype data for CD8⁺ cells specific for particular peptide containing

tetramers.

Figure 7 A-7C, inclusive, depict the results of experiments designed to compare fine specificities

of unsorted TILN populations.

Figure 8 summarizes data on stimulation of cytokine production by polyclonal and monoclonal

cell populations.

Figure 9 summarizes results of experiments designed to test PBL activity in the presence of NK

receptor specific monoclonal antibodies. Example 1

In order to make the desired tetramers, it was first necessary to prepare constructs which

would encode modified HLA-A^*0201 molecules. To do this, total RNA was extracted from

HLA-A^*0201 positive cells, and HLA-A^*0201 was then cloned, using specific primers for the

molecule, and reverse transcription polymerase chain reaction (RT-PCR). Altman et al, Science

274: 94-96 (October 4, 1996) but with a new 3' primer, i.e. 5'-

GCAGGATCCCGGCTCCCATCCTCA GGGTGAGGGGC-3' (SEQ ID NO: 48) incorporated

by reference, was followed. Simultaneously with the RT-PCR, the amino terminal nucleotide

sequence was altered to optimize protein expression in the vector used. See Garboczi et al,

Proc. Natl. Acad. Sci. USA 89: 3429 (1992) incorporated by reference. Once this was done, the

extracellular coding portion of the molecule was amplified, again using specific primers. The

resulting construct was recloned into a vector which would produce a BirA biotinylation

recognition site in frame at the 3 '-end of the HLA-A'0201 heavy chain. The modified HLA-

A^*0201 and B2 microglobulin were overexpressed in separate E. coli cultures. The resulting

inclusion bodies were purified and the HLA and B2 microglobulin recombinant proteins were

solubilized into urea, and then refolded, in a refolding solution, at 4 °C to form complexes. (The

refolding solution contained 100 mM Tris, at pH 8.0, L-arginine, 400 mM, EDTA, 2 mM,

reduced glutathione, 5 mM, oxidized glutathione, 0.5 mM, PMSF, 0.1 mM, HLA heavy chain,

and β2 microglobulin 1 μM, and 10 μM of the peptide of interest). The refolding solution was

concentrated to 7.5 ml, using standard techniques. Then, refolding buffer was exchanged with

BirA reaction buffer (Tris 100 mM, pH 7.5, NaCl 200 mM, Mg Cl₂ 5mM, PMSF 100 μM,

leupeptin 1 μM, and pepstatin 1 μM), the last three being added immediately before use. The complexes were then biotinylated with biotin holoenzyme synthase (the BirA

enzyme) by combining the refold mix containing the HLA-A2 complex with 50 μM enzyme,

100 mM biotin in 200 mM Tris, and 100 mM adenosine triphosphate. The mixture was

incubated overnight at room temperature. The biotinylated complexes were then purified, and

combined with phycoerythrin-labelled streptavidin, to produce tetrameric structures. These were

isolated, and reconstituted in small volumes, at a concentration of 1 mg/ml.

Example 2

Following isolation and reconstitution of the tetramers, these were tested to determine

if they could identify CTLs which express the cognate T cell receptor. To test this, monoclonal

CTL populations specific for SEQ ID NOS: 2, 4 and 5 (see infra), were used. In these

experiments, tetramers were prepared, using SEQ ID NOS: 3, 4 and 5 (see infra, but not SEQ

ID NO: 2) in a two color fluorescence assay with anti-CD8 antibodies, labelled with FITC. The

tetramers uniformly stained the CTLs, providing a signal two orders above background. The

specificity of CTL clone staining correlated well with the specificity of lytic activity displayed

by the same clones against ⁵¹Cr labelled T2 cells sensitized with SEQ ID NO: 2, 3, or 4, infra-

Example 3

The peptides, which were used in examples 1 and 2, i.e.:

Glu Ala Ala Gly He Gly He Leu Thr Val (SEQ ID NO: 2)

Glu Leu Ala Gly He Gly He Leu Thr Val (SEQ ID NO: 3)

Tyr Met Asp Gly Thr Met Ser Gin Val (SEQ ID NO: 4) and

Gly He Leu Gly Phe Val Phe Thr Leu (SEQ ID NO: 5) are all known to bind to HLA-A2 molecules, and stimulate CTLs. SEQ ID NO: 2 is derived

from Melan-A; SEQ ID NO: 3 is an analogue of SEQ ID NO: 2; SEQ ID NO: 4 is derived from

tyrosinase 2 (Skipper et al., J. Exp. Med. 183: 527 (1996)); SEQ ID NO: 5 is a control peptide

from influenza matrix protein (Gotch et al., 1987, Nature 326; 881). Hence, they were used in

the formation of the tetramers discussed supra. The tetramers were then used to detect specific

T lymphocytes in tumor infiltrated lymph nodes ("TILNS" hereafter).

TILNS were surgically removed from ten melanoma patients, who had been typed as

HLA- A^* 0201 positive. The TILNS were treated in the same manner as the TILNS in Romero

etal.,J. Immunol. 159: 2366-2374 (1997), incorporated by reference. Control lymph nodes from

normal lymph nodes ("NLNs" hereafter) were also prepared. The TILNs were treated to form

single cell suspensions, and were either not cultured, or cultured for anywhere from 1 -22 days

for expansion of specific T lymphocytes in medium supplemented with rIL-2 and rIL-7, and then

samples containing from 0.5-l.OxlO⁶ were stained with fluorescent, anti-CD3, anti-CD8

antibodies, as well as the tetramers described supra.

The anti-CD3 antibody was labelled with PerCP (peridinin chlorophyll protein). The

anti-CD8 antibody was labelled with FITC. A total of 1-2 ug of the tetramer being tested, in 20

ul volumes of PBS, 0.02 mM, 0.2% bovine serum albumin, 0.2% sodium azide, was used. It was

combined with the cells, for 30 minutes, at 4-8 °C. The cells were washed once, with 3 ml of the

same buffer, and were then analyzed immediately via flow cytometry.

A surprising high number of lymphocytes (0.12 to 21% of CD3⁺ and CD8⁺ lymph node

cells) specific for complexes of SEQ ID NO: 2 or 3 were found, in 11 of the 12 TILN samples

tested. The binding lymphocytes were CD3⁺ CD8⁺ lymph node cells. There were also high numbers of TILN cells which bound to the tetramers containing SEQ ID NO: 4, ranging from

0.14 to 0.63%.

Example 4

The samples discussed supra were analyzed further, emphasis being placed on TILN

samples which had been incubated overnight (5 samples) or stained immediately (2 samples).

Of these, the levels ofSEQ ID NO: 3 tetramer binding lymphocytes ranged from 0.12 to 1.29%.

There were no SEQ ID NO: 4/HLA-A2 positive TILNs found in these samples which had been

cultured overnight. This suggests that highest levels of positive lymphocytes are achieved after

short term in vitro culture.

Example 5

In order to confirm the suggestion outlined supra, cell suspensions were prepared from

a heavily tumor infiltrated lymph node and a normal lymph node ("NLN") from the same patient.

The suspensions were cultured, as described supra, overnight, for 15 days, or for 21 days and

analyzed for tetramer staining as described supra. Tetramer positive cells rose from 0.12 to

12.2% in the TILNs, but from 0.14 to 0.81 % in the NLNs when the 21 day cultured lymph node

cells were used. What was more drastic, however, was the absolute number of tetramer positive

cells. Specifically, the apparent number of SEQ ID NO: 3/HLA-A2 tetramer positive CD3^"

CD8⁺ cells rose from 210 to about 1.8x10⁵ after 15 days, and 5.38x 10⁵ after 21 days, representing

861 fold and 2562 fold expansions. Minor but significant expansion was seen in the NLNs, but

at two orders of magnitude less. Example 6

Studies were carried out to determine relationships between those TILN lymphocytes

which bind the tetramers, and CTL effectors capable of killing tumor cells expressing Melan A

antigen and to obtain highly homogenous populations of such CTL effectors.

To do so, TILN cells from a patient were cultured, for 21 days, as described supra, using

recombinant IL-2 and IL-7, and were stained with anti-CD3, labelled withperidinin chlorophyll

protein, and anti-CD8, labelled with FITC. These were then sorted into Melan A HLA-A2

tetramer positive, and negative cells, as described supra. Chromium release assays (Romero et

al. 1997, J. Immunol. ]_59: 2366) were carried out which showed that only the Melan- A/HLA-A2

tetramer positive T lymphocytes were able to kill T2 target cells sensitized with the peptide of

SEQ ID NO: 3.

Example 7

To further characterize the functional antigen specificity, TILNs which were sorted as

in example 5 were expanded for two weeks via non-specific stimulation with mitogen PHA, and

allogeneic, irradiated feeder cells. The sorted, expanded TILNs in a ⁵¹Cr release assay as

described supra against non-melanoma HLA-A2 expressing T2 cells, which had been sensitized

with one of seven peptides known to bind to HLA-A2 molecules, and to generate CTLs. The

assay was carried out under the conditions in example 6, supra.

It was found that lymphocytes which were specified to the tetramers containing the

decapeptide of SEQ ID NO: 3 only recognized this peptide. In contrast, tetramer negative cells

did not recognize any of the peptides. It was also observed that the Melan-A specific cells did in fact kill the autologous tumor

cell line efficiently.

Example 8

Apart from the importance of being able to visualize antigen specific CTLs, as is shown

supra, it is important to determine if these CTLs have encountered the antigen in vivo. Prior

work has established that expression of several antigens, including CD25, CD69, and CD45RO

are associated with T cell activation in CD8⁺ lymphocytes. These first two markers are

expressed transiently, while CD45RO is expressed in a more stable manner. Hence, the staining

method described supra was carried out, using anti-CD8 antibodies labelled with peridin

chlorophyll protein, anti-CD45RO, labelled with FITC, and the tetramers described supra. Ex

vivo cells suspensions were prepared from two patients, and then studied, with a tumor

infiltrated lymph node being tested with an unfiltrated node used for comparison.

It was found that the tetramer positive lymphocytes were predominantly, if not

exclusively in the CD45RO⁺, CD8⁺ lymphocytes. This indicated that such lymphocytes had

encountered antigen m vivo and were therefore being activated at the tumor site.

Example 9

In view of the experiments described supra, additional work was carried out to determine

if the tetramers could be used to isolated complex specific CTLs from circulating lymphocytes.

To do this, cryopreserved peripheral blood lymphocytes were thawed (37°C, 3 minutes), then

washed twice with standard culture medium. Magnetic cell sorting was carried out to select

CD8^T cells, following standard techniques. The cells were then analyzed, using CD45RA^" specific antibodies labelled with cytochrome, and CD28 labelled with FITC, following the

protocols of example 8. A total of 0.11% of the CD8⁺ PBLs were tetramer positive and, most

notably, most were CD28⁺. Approximately 44% of the CD8", tetramer specific cells, were

CD45RA\ This shows that uncultured, unstimulated blood lymphocytes can be assayed for

appropriate CTLs specific for the complex of interest.

Example 10

Further analysis were carried out with the tetramers of SEQ ID NO: 4. These stained

CTLs known to be specific for complexes of SEQ ID NO: 4 and HLA-A2, but not others. The

tetramer was then used to study PBMCs of individuals who were not known to be suffering from

influenza infection, to determine if the tetramers could be used to assay for low frequency CTLs,

like memory cells, ex vivo. The assay was carried out, as described in examples 8 and 9, supra.

The percentage of complex specific CD8⁺ cells was considerably higher than the

percentage estimated previously using limiting dilution analysis, but are similar to the percentage

observed, using ELISPOT analysis, as per Lalvani et al., J. Exp. Med. 186: 859-865 (1997),

which is incorporated by reference.

Example 11

Studies were then carried out to phenotype the complex specific, CD8⁺ cells. A large

number of PBMCs were obtained from an individual and, using the assay of example 10, supra,

the frequency of CD8⁺ cells specific for the complexes was consistently found to be about

1/3700 of the PBMCs, and 1/2500 of small lymphocytes which are activated, or "memory" CD8"

cells. When tested for cell surface markers, these were found to be CD45RO7CD45RA , consistent with such cells having previously encountered antigen. At least 74% of the positive

CTLs expressed the VB17 chain, as determined by cell surface staining, as compared to 4% of

the negatives, which confirms results shown by Lehner et al., J. Exp. Med. 181: 79-91 (1995).

Only 36%o of the positive cells were CD28⁺, compared to 75% of the negative CTLs. Again, this

is consistent with the majority of the cells having reduced proliferative potential due to longer

replicative history. This shows that, using the triple staining methodology, a very pure cell

population can be obtained, by applying, e.g., FACS on peripheral blood lymphocytes, thereby

identifying CTLs.

Example 12

Specificity and activation of the CD8⁺ PBMCs discussed supra was studied using the

ELISpot technique, referred to supra. This technique yields spots of color whenever a CTL

releases gamma interferon. In control experiments, a CTL clone specific for SEQ ID NO:

4 HLA-A2 complexes was stained with the tetramer and anti-CD8. Then, 100 double positive

cells were sorted directly into duplicate wells of ELISPOT plates containing T2 cells, which had

been prepared with either SEQ ID NO: 5, or as a control, SEQ ID NO: 2. No gamma interferon

was released from controls, indicating that tetramer staining did not per se activate the cells. The

SEQ ID NO: 5 containing matrices, however, provoked gamma interferon release corresponding

to about one spot per CTL per day, indicating all cells had been activated by their cognate

peptide.

Following this, the PBMCs were stained, sorted into sets of 15 cells and duplicate

ELISPOT wells for 1 or 2 days of incubation with peptide pulsed targets. Interferon release was

observed only when the peptide of SEQ ID NO: 5 was used. Pulsing with other peptides did not result in gamma interferon release, nor did tetramer negative CD8⁺ cells release IFN-gamma on

exposure to any peptide.

Since the number of spots formed was close to the number of PBMCs in the wells, this

suggests that the majority of the tetramer positive, CD8⁺ cells were capable of responding within

one day, thus supporting the concept that most circulating memory CD8⁺ cells are capable of

rapid effector function when triggered by exposure to cognate peptide in the absence of cytokine,

even when the presenting cells are not professional antigen presenting cells.

Example 13

Studies were then carried out to confirm the specificity of tetramer specific, CD8"

specific PBMCs, and to study proliferative potential. To do this, single PBMCs were sorted

directly into cloning wells. After two weeks, 15 of 60 wells seeded with these cells contained

proliferating blasts. The four most proliferative clones were tested in duplicate ELISPOT assays

for specificity, along with one proliferating CD 8^', tetramer negative clone. What was found was

that all four of the CD8⁺, tetramer positive clones synthesized gamma interferon when exposed

to targets pulsed with SEQ ID NO: 4, but not when pulsed with a control peptide. The tetramer

negative, CD8⁺ cells failed to respond at all. Further, one of the clones was tested in a chromium

release assay, where it specifically lysed target cells pulsed with SEQ ID NO: 5.

Example 14

The potential of tetramer based, FACS sorting in analyzing tumor specific CTLs

responses was studied using tetramers containing SEQ ID NO: 2, as described supra. The

tetramer was then used to stain and sort a polyclonal CTL line, which had been generated from an HLA-A2", melanoma infiltrated lymph node. Only a small percentage of cells (6%), could

be double stained with the specific tetramers and antibody to CD8. The polyclonal line did kill

A2 matched targets pulsed with SEQ ID NO: 2, although it was poor at killing A2 matched

melanoma lines expressing Melan-A, and demonstrated small background killing of target cells

pulsed with a negative control.

An aliquot of the polyclonal line was then enriched for tetramer specific, CD8 cells by

FACS sorting. This line killed melanoma lines and target pulsed with SEQ ID NO: 2 much more

efficiently than the original polyclonal line, and showed no background killing. The specific

cytotoxicity correlated entirely with the percentage which stained tetramers, indicating that

tetramer staining, followed by FACS sorting, can be used to generate highly effective,

tumoricidal, CTL lines.

Example 15

These experiments describe in vitro expansion of Melan A specific precursors from

PBMCs of melanoma patients.

Samples of PBMCs were taken from seven melanoma patients, using standard

techniques, and were enriched for CD8⁺ lymphocytes via magnetic cell sorting, to give

populations of cells which were over 99% CD37CD87

These cells were then cultured, overnight, under standard conditions, and were then

stained with tetramers of HLA-A* 0201 and SEQ ID NO: 3, which were labelled with

phycoerythrin, and anti-CD8 monoclonal antibodies (Staining was accomplished by combining

the cells with 20μl of PBS with 2% FCS, and incubating for 15 minutes at 4°C, after which 20μl

of anti-CD8 monoclonal antibodies labelled with FITC were added, and incubated for 30 minutes at 4°C). Cells were then washed, once, with the buffer described supra, and analyzed

by standard flow cytometry.

The frequency of tetramer positive cells ranged from 0.03% to 0.09% of the total CD8⁺

population. These frequencies were close to the detection limit of the flow cytometer, which led

to the next experiments.

Example 16

In view of the low levels of positive cells found, PBMCs were taken from 3 of the 7

melanoma patients described supra, and were stimulated three times, at weekly intervals, with

either SEQ ID NO: 2 or SEQ ID NO: 3 pulsed, autologous PBMC cells. To elaborate, the

samples were stimulated, initially, by adding 1 μM of peptide directly into the medium. Cultures

were then stimulated weekly by adding autologous PBMCs (3x10⁶ cells/well), which had been

pulsed for 2 hours at 37°C, in serum free medium, with lμM of the peptide of SEQ ID NO:2 or

SEQ ID NO: 3 and 3μg/ml of human β2 microglobulin. These peptide pulsed cells were washed

extensively, irradiated (3000 rad), and adjusted to an appropriate volume before being added to

the responder cell population. Both IL-2 ( 1 OU/ml), and IL-7 ( 1 Ong/ml), were added during the

first two stimulation cycles, and IL-2 alone was added thereafter, at 100 U/ml.

The assay described in example 15 was carried out on the samples 7 days after each

stimulation cycle.

During the first week of stimulation, the total number of tetramer positive cells declined

consistently, increased moderately after the second stimulation, and significantly after the third

one. Stimulation with the peptide of SEQ ID NO: 3 resulted in more vigorous expansion than

did stimulation with SEQ ID NO: 2. The difference in stimulatory capacity was apparent after the first cycle and became very marked after the third one. The results are summarized in the

following table.

Example 17

Cultures were tested seven days after the third stimulation cycle for their capacity to lyse

T2 cells, in the absence or presence of each of the two peptides. (T2 cells are known as cells

which are HLA- A* 0201 positive, are defective in antigen processing ability, but present

exogenously supplied peptides effectively). The assays were carried out by labelling the T2 cells

with ⁵¹Cr for 1 hour at 37°C, followed by two washes. Labelled target cells (1000 cells in 5 Oμl)

were then added to varying amounts of effector cells (in 50μl) in V-bottom microwells, either

in the presence of 1 μg/ml of the peptide of SEQ ID NO: 2, 1 μg/ml of the peptide of SEQ ID NO:

3, or no peptide.

The effector cells had been preincubated for at least 20 minutes at 37 °C in the presence

of unlabelled K562 cells (50,000/well) to eliminate non-specific lysis due to NK-like effectors

in the stimulated T cell populations. Chromium release was measured after 4 hours. The results are presented in figure 1. Frequency of HLA-A*0201/SEQ ID NO: 3

tetramer positive cells detected in the cultures correlated directly with peptide specific

cytotoxicity. Also, CTLs which were generated following stimulation with SEQ ID NO: 3 cross

recognized SEQ ID NO: 2.

Example 18

Additional experiments were carried out to document the specificity of CD8⁺, HLA- A*

0201 /peptide tetramer positive lymphocytes which had been generated supra.

In these experiments, PBMCs from one of the three patients described supra were

stimulated with autologous PBMCs that had been pulsed with SEQ ID NO: 1, SEQ ID NO: 2,

SEQ ID NO: 3, the peptide Ala Ala Ala Gly He Gly He Leu Thr Val (SEQ ID NO: 37), Ala Leu

Ala Gly He Gly He Leu Thr Val (SEQ ID NO: 38), or an immunodominant HLA-A2 restricted

epitope from influenza matrix protein, presented as SEQ ID NO: 5, supra. (SEQ ID NOS: 37

and 38 are Melan A analogues, and hence are related to SEQ ID NOS: 1,2 and 3). The

stimulation protocol was identical to that given in example 17, supra.

A small number of HLA- A* 0201/ peptide tetramer positive cells were detected at the

end of the first stimulation cycle in the cultures stimulated using the influenza peptide, and was

equivalent to the number of such cells detected in uncultured cells. The number declined after

a second round of stimulation.

Expansion of the tetramer positive lymphocytes occurred with both SEQ ID NOS : 1 and

2, directly demonstrating the ability of the peptides to stimulate in vitro expansion of Melan-A

specific lymphocytes in peripheral blood. Their stimulatory capacity appears to be similar. Notwithstanding this point, the increase in proportion of tetramer positive lymphocytes

wasmuchmorepronouncedwhenanalogues ofSEQIDNO: 1 were used. The largest expansion

occurred with SEQ ID NO: 38. The following table summarizes these results. The values show

the percentage of cells which were positive both for the tetramer discussed supra, and the FITC

labelled, anti CD8 antibody.

Peptide

Example 19

Following the work in example 18, the viable cells recovered after the third stimulation

were titrated as effectors in a ⁵¹Cr release assay of the type described supra.

Frequency of tetramer positive cells correlated directly with the level of specific

cytotoxicity, and CTLs resulting from the analogue driven expansion effectively lysed target

cells sensitized with SEQ ID NO: 2.

Example 20

Examples 15-19 show that it is possible to directly visualize antigen specific T

lymphocytes using flourescent tetramers, and to separate them from bulk cultures. They suggest the possibility of early stage, in vitro stimulation with peptide, and continued expansion in

antigen independent fashion. This was tested in the experiments which follow.

Samples of CD8⁺, tetramer positive cells were isolated from each of the cultures

described in example 18, seven days after the second cycle of stimulation. These sorted cells

were then expanded in vitro by stimulation with phytohemagglutinin, and the resulting cells were

tested for their lytic activity on T2 cells, in the same manner described supra, in the absence or

presence of the peptide used in the particular culture from which the cells were taken originally,

or SEQ ID NO: 2. Concurrently, the ability of these CD8⁺ cells to lyse tumor cells was tested,

using Melan-A positive tumor cell line Me 290 as the target.

The results, presented in figure 2, show that all the cell populations exhibited a high

degree of specific lysis against target cells which had been pulsed with the corresponding

stimulatory peptide and with cells pulsed with SEQ ID NO: 2. They also exhibited high

tumoricidal activity. Me 290 is illustrative of additional Melan-A positive tumor cell lines which

were lysed. Melan-A negative tumor cell lines were not lysed. Effectoπtarget ratios necessary

for achieving half maximal lysis ranged from 3:1 to 15:1.

Example 21

The relative avidity of peptide antigen recognition of the different cell populations which

had been obtained was assessed in CTL assays, following standard protocols. Specifically, 1000

target cells (50μls) were incubated with varying concentrations of peptides (50μl) for 15 minutes

at room temperature before effector cells were added. The concentration of each peptide

required to achieve 50% maximal lysis of target was determined, using standard methods. To

make comparison easier, relative activity of peptide was calculated as the 50% concentration of SEQ ID NO: 1, divided by the 50% concentration of the tested peptide. Figure 3 shows these

results. These results are also presented in the table which follows. The values are the peptide

concentrations required for 50% maximal activity.

The relative avidity of the different lines for SEQ ID NO: 1 and SEQ ID NO: 2 were

remarkably similar, except for the line obtained using SEQ ID NO: 39, which recognized SEQ

ID NO: 1 about 3 fold less efficiently, and SEQ ID NO: 2 5-7 fold less efficiently as compared

to other lines.

The results are also presented in terms of relative antigenic activity. Note that, regardless

of the peptide used for in vitro expansion, all lines recognized SEQ ID NO: 2 more efficiently

than SEQ ID NO: 1. Analogues were recognized more efficiently than both parental sequences

by all lines, notwithstanding some differences in relative antigenicity.

Preference of a line for a certain analogue did not correlate with the analogue used to

generate the line: in all cases, SEQ ID NO: 38 was the peptide most efficiently recognized.

Culture stimulated with:

Melan-A Melan-A E26A A27L E26A/A27

27-35 26-35 L

(SEQ ID (SEQ ID (SEQ ID (SEQ ID (SEQ ID NO: 1) NO: 2) NO: 37) NO: 3) NO: 38)

Peptide [nM] 50%:

Melan-A 27-35 (SEQ ID NO: 1) 27 22 18 37 65

Melan-A 26-35 (SEQ ID NO: 2) 3 3 5 5 20

E26A (SEQ ID NO: 37) 0.09 0.14 0.12 0.19 1 Melan-A Melan-A E26A A27L E26A/A27

27-35 26-35 L

A27L (SEQ ID NO: 3) 0.15 0.2 0.2 0.16 0.4

E26A/A27L (SEQ ID NO: 38) 0.02 0.01 0.008 0.005 0.5

Relative antigenic activity:

Melan-A 27-35 (SEQ ID NO: 1) 1 1 1 1 1 Melan-A 26-35 (SEQ ID NO: 2) 9 7 4 7 3

E26A (SEQ ID NO: 37) 300 157 183 194 65 A27L (SEQ ID NO: 3) 180 110 110 231 162 E26A/A27L (SEQ ID NO: 38) 1350 2200 2750 7400 130

Example 22

The peptide presented supra as SEQ ID NO: 4 is derived from the tyrosinase molecule,

which is known to function as a tumor rejection antigen precursor, with the peptide of SEQ ID

NO: 4 functioning as a tumor rejection antigen. See, e.g., Brichard, et al., J. Exp. Med. 178:489

(1993); Wδlfel, et al., Eur. J. Immunol. 24:759 (1994). Work by others has focused on study of

the binding properties of an analog of this peptide, i.e., one where asparagine at position 3 of the

peptide, has been replaced by aspartic acid. (See SEQ ID NO: 22, supra). This modification

occurs naturally, via post transiational processes. Skipper, et al., J. Exp. Med. 183:527 (1996)

have shown that this difference significantly affects T cell recognition. The experiments which

follow are directed to a determination of what role single amino acid side chains contribute to

the interaction between SEQ ID NO: 4 and HLA-A* 0201 molecules.

First, single alanine substituted derivatives of SEQ ID NO: 4 were prepared, i.e.: Ala Met Asp Gly Thr Met Ser Gin Val (SEQ ID NO: 39)

Tyr Ala Asp Gly Thr Met Ser Gin Val (SEQ ID NO: 40)

Tyr Met Ala Gly Thr Met Ser Gin Val (SEQ ID NO: 41)

Tyr Met Asp Ala Thr Met Ser Gin Val (SEQ ID NO: 42)

Tyr Met Asp Gly Ala Met Ser Gin Val (SEQ ID NO: 43)

Tyr Met Asp Gly Thr Ala Ser Gin Val (SEQ ID NO: 44)

Tyr Met Asp Gly Thr Met Ala Gin Val (SEQ ID NO: 45)

Tyr Met Asp Gly Thr Met Ser Ala Val (SEQ ID NO: 46)

Tyr Met Asp Gly Thr Met Ser Gin Ala (SEQ ID NO: 47)

These peptides were synthesized using standard methodologies, and were then tested in

a functional competition assay. To elaborate, the assumption was that these peptides would bind

to HLA-A*0201 molecules. It is known that SEQ ID NO: 1 binds to HLA-A*0201. Hence, T2

cells, described supra, were labelled with ⁵¹ Cr in the presence of a well known monoclonal

antibody, i.e., W6/32, which binds MHC - Class I molecules.

Then, varying concentrations of test peptides SEQ ID NOS: 39-47 were added, in 50 μl

aliquots, to the labelled T2 cells (1000 cells/well), for 15 minutes at room temperature, after

which a suboptimal dose of SEQ ID NO: 1 (125nM in 50 μl), was added, together with 5000

cells/well of an HLA-A*0201/SEQ ID NO: 1 restricted CTL clone, as taught by Valmori, et al.,

J. Immunol 160: 1750 (1998). Chromium release was measured after 4 hours of incubation at

37 °C. The concentration of each competitor peptide required to achieve 50% inhibitions of target

cell lysis was determined. This is indicated in the Table which follows, as [nM] 50%. The

comparison is facilitated by presenting relative competitor activity as [nM] 50% of SEQ ID NO: 4. This peptide' s activity is assigned arbitrary competitor activity of 1. As a control, an

influenza matrix peptide was used.

It was found that SEQ ID NO: 4 displays intermediate competitor activity, i.e., to secure

50%) of maximal lysis, a concentration of lOOnM was needed, as compared to 6nM for SEQ ID

NO: 5, known to be a high affinity binder. It was surprising that changes at putative anchor

positions 2 and 9 had marginal impact. Changes at positions 7 and 8, for example, had a much

greater impact.

IΛJD 5545.1-JEL/NDH

Table

Binding of single A substituted Tyrosinase 368-376 peptide variants to HLA-A* 0201

Peptide Sequence Competitor Relative

activity competitor

[nM] 50% activity

G

Tyrosinase 368-376 YMDGTMSQV (SEQ ID NO: 4) 102

CZ5

1 H— H analogues AMDGTMSQV (SEQ ID NO: 39) 200 0.5

H YADGTMSQV (SEQ ID NO: 40) 300 0.3

H

W s YMAGTMSQV (SEQ ID NO: 41) 90 1.1

X YMDATMSQV (SEQ ID NO: 42) 100 1.0 w m YMDGAMSQV (SEQ ID NO: 43) 80 1.3

H YMDGTASQV (SEQ ID NO: 44) 200 0.6 d YMDGTMAQV (SEQ ID NO: 45) 50 r 2.0 w YMDGTMSAV (SEQ ID NO: 46) 30 3.3 t YMDGTMSQA (SEQ ID NO: 47) 80 1.3

Influenza A matrix 58-66 GILGFVFTL (SEQ ID NO: 5) 17

Example 23

It had been observed previously that complexes formed between HLA- A* 0201 molecules

and SEQ ID NO: 2 are unstable, dissociating completely in under one hour when incubated at

37°C. Stability studies were carried out on complexes of HLA-A*0201 and SEQ ID NO: 4. In

brief, T2 cells were loaded with saturating concentrations (lOμM) of one of SEQ ID NO: 4,

SEQ ID NO: 2, or control peptide SEQ ID NO: 5. Following overnight incubation with the

peptides and β2-micro-globulin (3μg/ml) in serum free medium, the T2 cells were treated with

ementine (10⁴ M) to inhibit protein synthesis, and incubated at 37 °C. The cells were stained,

at different time periods, by incubating an aliquot with anti-HLA-A2 specific monoclonal

antibody BB7.2, to measure HLA-A2 antigen expression. The results were expressed as relative

complex stability, defined by the formula:

[Mean fluorescence test peptide - background mean fluorescence] [Mean fluorescence SEQ ID NO: 48 - background mean fluorescence]

"Background" refers to fluorescence values obtained on T2 cells treated similarly but for absence

of exogenous peptide.

Figure 4 presents these results. They confirm the instability of HLA-A2/SEQ ID NO:

1 complexes, and show that the complexes of HLA-A2 and SEQ ID NO: 4 are stable over a 6

hour period.

Example 24

It has been observed, by Romero, et al, J. Exp. Med. 188:1641 (1998) incorporated by

reference, and in the preceding examples, that significant numbers of HLA-A2/SEQ ID NO: 4 positive, CD8⁺ T lymphocytes can be detected in short term cultured TILNs from HLA-A*0201

positive melanoma patients. Studies were carried out to assess the antigenic and functional

specificity of such positive lymphocytes.

In brief, positive and negative fractions were separated via flow cytometry sorting, using

tetramers of HLA-A2/SEQ ID NO: 4, and CD8 specific monoclonal antibodies, as described

supra. Following separation, cells were expanded by stimulation with PHA and irradiated

allogeneic PBMCs, as described supra, over a two week period. The fractions were tested for

their ability to lyse ⁵¹Cr labelled T2 cells in the absence or presence of 4μM of SEQ ID NO: 4.

The CD8^~, tetramer positive fraction lysed the T2 cells in the presence of SEQ ID NO: 4, but not

its absence. Negative fractions did not lyse the target.

Example 25

These experiments were designed to obtain monoclonal T cell populations specific for

SEQ ID NO: 4 containing complexes. To do this, limiting dilution cultures were set up

immediately after the cell sorting described in example 24. Limiting dilution was carried out

following Valmori, et al., J. Immunol. 160:1750 (1998), incorporated by reference, using

irradiated allogeneic PBMCs, EBV transformed B lymphocytes, PHA and recombinant IL-2.

Six specific CTL clones were secured, and a seventh clone was obtained from

unfractionated TILNs after in vitro stimulation in an independent cloning experiment. These

constitute monoclonal T cell populations.

Example 26 These experiments were designed to analyze both polyclonal, monospecific T cell

populations, and the monoclonal T cell populations described supra, for tumor antigen

recognition, avidity, and fine specificity. This example describes the antigen recognition

experiments.

Three cell lines were used. NA8-MEL is an HLA-A*0201 positive melanoma cell line

which does not express tyrosinase. NA8-MEL tyr^" is a cell line derived by transfecting NA8

with cDNA encoding tyrosinase. Cell line Me 290 described supra as an HLA-A* 0201 positive,

Melan-A⁺ cell line is also tyrosinase⁺.

Samples of each of these lines were combined with one of a polyclonal, monospecific

CTL population obtained by using the tetramers of SEQ ID NO: 4 and HLA-A* 0201, supra, or

the monoclonal T cell populations described supra.

The melanoma cell lines and T cells were combined, at varying ratios, either with or

without the peptide, at lμM concentration. The tumor cells had been labelled with ⁵¹Cr prior to

combining them with the T cells. ⁵¹Cr release was measured after 4 hours of incubation, using

the same parameters as are described supra.

These results are set out in figure 5. The panel labelled Tyr. tetr⁺ shows data from the

polyclonal cell population. The others are the monoclonal lines described supra.

All of the clones and the polyclonal population lysed the NA8-MEL cell line, which is

tyrosinase negative, only when the peptide of SEQ ID NO: 4 was added. The other lines, i.e.,

Me 290 and NA8-MELtyr+ were lysed equally well, both in the presence of the peptide and

without it.

Example 27 Fine antigenic specificity of the T cell populations discussed in example 26 was analyzed,

by quantitating the relative antigenic activity of the variants set out in SEQ ID NOS: 39-47 in

standard CTL assays. These were carried out by using a ^5ICr release assay, as described supra.

Relative avidity was determined by comparison to the concentration of the peptide of SEQ ID

NO: 4 required to obtain half maximal lysis at an effector: target ratio of 10:1. The results are

presented in the following Table.

LUD 5545.1-JEL/NDH

LUD 5545.1-JEL/NDH

45 YMDGTMAQV O.0004 <0.003 <0.0004 0.04 O.003 <0.02 <0.00 <0.015

2

46 YMDGTMSAV 0.05 <0.003 O.0004 1 0.05 <0.02 <0.00 100

2

47 YMDGTMSQA 0.06 O.003 0.05 1 0.5 0.5 0.2 0.03

GO c

GO

H

H Cl H m w

U1

GO

X w

H

Each clone displayed unique fine specificity. Replacement of residues at positions 2 and

9 did not drastically affect antigen recognition in the majority of cases, which is consistent with

lack of direct action between side chains located at these positions, and T cell receptors, as

suggested by Madden, et al., Cell 75:693 (1993). The results indicate that position 7 of the

peptide (a Ser residue), was important, while replacement at different positions gave highly

variable results. The net result is a showing of wide heterogeneity in the recognition of epitopes

by CTLs specific to SEQ ID NO: 4, and derived from a single, infiltrated lymph node.

Example 28

Example 15, supra, showed how tetramers of HLA-A* 0201 and a Melan-A derived

peptide could be used to measure the frequency of CD8⁺ T lymphocytes in a population. Similar

experiments were carried out using tetramers of HLA-A* 0201 and SEQ ID NO: 4.

The same protocol as was used in example 15, but for a different tetramer, was carried

out on highly enriched CD8⁺ T cells from PBMCs of 10 HLA-A* 0201 positive melanoma

patients. Seven samples showed frequencies of tetramer⁺ cells ranging from less than 0.01 % and

up to 0.03%). These ranges were not very much above the level of detection afforded by the flow

cytometry assay, so the enriched, CD8⁺ T lymphocytes for all 10 samples were stimulated with

SEQ ID NO: 4, using the protocol of example 16. supra, with minor modifications. Specifically,

cells were initially stimulated with 100 μM of peptide directly into the culture medium, and

weekly stimulations were carried out with additional lOOμM samples of peptide. Additionally,

IL-2 (100 U/ml) and IL-7 (10 ng/ml) were added during the first stimulation cycle, and IL-2

alone (100 U/ml) thereafter. Cultures were monitored seven days after stimulation for the

presence of tetramer positive cells. The results are summarized in the following Table. What is interesting is that positive

lymphocytes were detectable in six out of 10 samples after only short term culture. (The other

four samples were positive, but at the limits of detection for the assay).

The values refer to the percentage of A2/SEQ ID NO:4 tetramer positive T cells, relative to the

total number of CD8⁺ lymphocytes.

Example 29

In additional experiments, tetramers of HLA-A *0201 molecules and SEQ ID NOS: 2,

4 and now SEQ ID NO: 9, i.e., MAGE-3 derived peptide FLWGPRALV, described supra, were

made, as described in the preceding examples. These tetramers were then used to stain CTLs,

as described in the previous examples. To review, cell samples were stained with PE labelled

tetramer for 15 minutes at 37°C, after which tricolor labelled, anti-CD8" antibodies were added,

on ice, for 15 minutes. Following this, the materials were washed extensively, and then analyzed

by FACS. The tetramers stained CTLs which had previously been identified as being specific for

complexes of HLA-A *0201 and the peptide used, but did not stain with other, non-specific

peptides.

Example 30

These experiments describe staining of different types of samples from different patients.

Five patients were typed for expression of tumor antigens, following standard

methodologies. Different types of samples were taken from these patients, all of whom had

histologically confirmed, malignant melanoma. The characteristics of the patients and samples

involved are set forth in the following table:

Patient L02 was leukapheresed, and PBLs from this patient were cryopreserved

immediately. These PBLS were analyzed after thawing and overnight culture in Iscove's medium

with 5% human serum added. Two additional PBL samples were taken over aperiod of 3 weeks

for comparative analysis.

In the case of the patient referred to as DECH, PBLs were separated from a blood sample,

and cryopreserved. Following thawing, these PBLs were pulsed with lOμM of SEQ ID NO. 9 in "CTL medium" (Iscove's medium, 5% human serum, IL-2 at 100 U/ml), plus IL-7 (lOng/ml).

After this, they were cultured in CTL medium for 2 weeks until they were analyzed.

Patient MM 14 and MM 18 had melanoma - infiltrated lymph nodes removed from the left

or right axillae, respectively, disrupted, and cultured in the CTL medium described supra, plus

IL-7 (lOng/ml). Any proliferating blasts were expanded in the CTL medium for 14 days (for

MM14), or 23 days (for MM18), before analysis.

In the case of MM 15 , a skin metastasis was removed from the left shoulder, mechanically

disrupted, and cultured in the medium described supra for 13 weeks. There was a small

population of TILs, which was stimulated with PHA (5μg/ml), and expanded in medium for two

more weeks before analysis.

Tetramer staining was carried out as described supra, using appropriate tetramers. In

other words, tetramers containing SEQ ID NO. 9 were used in MAGE-3 positive samples from

DECH. Melan-A positive patient samples were screened with tetramers containing SEQ ID NO.

3, while tyrosinase positive patient staples were screened with tetramers containing SEQ ID NO.

4.

Tissue culture of tumor infiltrated lymph nodes of MM 14 revealed CD8^{^} cells specific

for both the tyrosinase and Melan-A peptides. The cultured LNs from MM18 were positive for

SEQ ID NO. 3 only. Patient L02 had readily detectable CD8⁺ cells specific to SEQ ID NO: 3

in peripheral blood at a frequency similar to the frequency of influenza specific CTLs in normal

patients, as described by Dunbar, et al., Curr. Biol 8:413 (1998). This frequency remained fairly

constant over the 3 week period of analysis. Patient DECH showed no CD8⁺ cells specific for

SEQ ID NO: 9 in PBLs on direct analysis; however, two weeks after a single pulse with the

peptide, a small population of cells were detected. Finally, patient MM15 showed very few CD8⁺ cells in a PHA stimulated culture from skin metastasis; however, a high portion of these

were stained by the tetramers containing SEQ ID NO. 3.

Example 31

The preceding example described how CD8⁺ cells were identified. In these experiments,

CTLs are cloned, directly, from these CD8⁺ populations.

To do this, single cells were sorted, directly, into U bottom, 96 well plates coated

previously with anti-CD3 and anti-CD28 antibodies (both at 100 ng/ml), in PBS, containing 10⁵

irradiated B cells in CTL medium, as described supra, plus IL-4 (20ng/ml). The anti-CD3

provides a solid phase mitogenic signal, and the anti-CD28 antibodies were provided in case the

cells were capable of costimulation through CD28.

The cloning plates were incubated at 37°C in 5% CO₂ for 10-14 days, without any

manipulation. Then, they were restimulated using PHA (5 μg/ml), together with irradiated,

allogenic PBLs and B cells. All clones were expanded to at least 5xl0⁶ cells, in order to permit

characterization and banking. Several of the clones were stimulated further, in order to test

proliferative potential, and the total potential yield was calculated from the number of CTLs

generated in culture, and the number banked.

The table within follows summarizes these results. Essentially, between 2 and 13% of

sorted clones (average: 6.5%) expanded sufficiently to cryopreserve. Additionally, about three

times as many proliferated to a lesser degree.

The expanded clones were then tested for antigen specificity using either tetramers as

described supra, with confirmance by a ⁵¹Cr release assay and ELISPOT analysis. With respect to the ⁵¹Cr release assay, this was carried out by T2 cells which had been

pulsed with relevant peptide (SEQ ID NOS: 3, 4 or 9), or a negative control, as well as cell lines

SK-Mel-29 and SK-Mel 23, both of which are HLA-A2.T, and express both Melan-A and

tyrosinase. The lysis percentage was calculated by the formula:

100 x (specific release - spontaneous release) (maximal release - spontaneous release)

Thirty three of thirty four of the clones were positive for their antigen, showing that the direct

cloning was extremely efficient, as compared to methods where the vast majority of cloned cells

do not have the desired specificity. Ten of these were tested by ⁵¹Cr release, and three of these

were tested via ELISPOT. All showed specificity for relevant peptide.

With respect to cloning efficiency, no clear differences were seen, although peptide

stimulated cultures required the greatest number of starting lymphocytes to generate clones:

Example 32 The expanded CTLs were then tested for surface phenotytpes via FACS, using FITC

labelled antibodies against CD62L (L-Selectin), CDl lb, CD44, CD45RO, CD45RA, TCRαβ,

and CD 49a-e., inclusive (integrin-Bl-6, VLA 1-6), followed by goat-antimouse Ig. Also, anti-

CLA (cutaneous lymphocyte antigen), and FITC labelled mouse-antirat IgM labelled with FITC,

were used.

All clones tested were CD8⁺, and TCRαβ^{^}, which was to be expected. They were also

all CD45RO", CD45RA \ CD44⁺ and CD62L^", which is consistent with previous antigenic

stimulation. The lack of CD62L on surfaces suggested that if these clones were infused during

immunotherapy, they would not home to lymph nodes via high endothelial venues, but would

traffic through peripheral circulation. The clones were CD1 lb and CD49d positive, which is

consistent with cells capable of migrating through activated vascular endothelium which express

VCAM-1 - a phenotype associated with metastatic melanoma (Rice, et al., Science 246:1303

(1989); Rice, et al., J. Exp. Med. 171 :1369 (1990)).

A minority of clones were CLA positive suggesting that clones could be selected which

would or would not home to skin following infusion. See Picker, et al., Nature 349:796 (1991);

Ogg, et al., J. Exp. Med. 188:1203 (1998). Hence, it may be possible to pre-select clones for

immunotherapy according to their homing markers, via triple staining for surface molecules prior

to cloning. Also, clones with the best proliferative potential might be pre-selected by only

sorting tetramer ⁺ cells which express markers, such as CD28.

Example 33

In this example, and the examples which follow, additional characterization studies were

carried out on both melanoma patients and healthy volunteers. Patients and donors were selected based upon HLA-A2 expression, determined by using

allele specific mAb BB7.2. PBMCs were isolated from samples, using standard methods. The

CD8⁺ cells in these samples were purified via two rounds of positive selection, using magnetic

beads, as described supra. These cells were greater than 98% CD3⁺CD8⁺.

These CD8⁺ cells were then stained with tetramers of HLA-A*0201 and SEQ ID NO: 3,

SEQ ID NO: 4 or control peptide SEQ ID NO: 5 and anti-CD8⁺ monoclonal antibodies labelled

with FITC and Cy-Chrome (a tandem flourescent conjugate system) (10⁶ cells added to 50 μl of

PBS containing the antibodies and the tetramers, as well as 2% azide and 2% BSA. The

incubation period was 40 minutes, at a temperature of 4 °C). Following the incubation, the cells

were washed, once, in the buffer used for the binding step, and then analyzed immediately via

FACS.

It was found that CD8⁺ cells specific to tetramers containing either SEQ ID NO: 3 or

control peptide SEQ ID NO: 5 were present in both melanoma patients and healthy subjects.

CD8⁺ cells specific to tetramers containing SEQ ID NO: 4 were not found. Note,

however, that these samples had not been stimulated with antigen in vitro. As was shown, supra.

even a short period of in vitro stimulation with antigen was sufficient to cause expansion of CD8^*

cells specific to SEQ ID NO: 4 to the point where they could be detected in 6 of 10 patients

tested.

Example 34

A series of experiments were then carried out to determine levels of non-specific, HLA-

A2 tetramer positive CD8⁺ cells in healthy, HLA-A2 negative patients. The same tetramers as

are described in example 33, supra, were used, and the same protocol for separation and analysis were used. These results are set out in the table which follows For compaπson. tests were also run on samples taken from HLA-A2 positive, healthy donors, HLA-A2 positive melanoma patients, and HLA-A2 positive patients who suffered from both melanoma and viti go.

Percentage of circulating HLA-A2 Melan-A_{26 3S} A27L tetramer and Flu-MA_{<8 66} tetramer^' in CD8~ T cells

The lower detection limit for CD8^* cells which reacted with tetramers of SEQ ID NO 3 was about 0.04%, while it was under 0.02% for CD8^~ cells positive for SEQ ID NO: 5 containing tetramers These limits of detection are lower than limits of detection for TILS, as shown m Romero, et al, J. Exp. Med. 188: 1641-1650 (1998), due to differences m the type of sample assayed. Of the 13 melanoma and melanoma/vitihgo patients tested, 10 showed significant numbers of HLA-A2/SEQ ID NO 3 positive CD8^* cells, while six of ten healthy HLA-A2 positive donors showed significant numbers as well Note that these CDS" cells were almost all naive cells, as compared to the mix of naive and antigen experienced cells found in

melanoma patients.

Example 35

Prior work has established that the CD45RA and CD45RO isoforms can be used to

identify naive and memory T cells, respectively. See Young, et al., Eur. J. Immunol. 27: 2383-

2390 (1997), incorporated by reference. It has also been shown that circulating, CD28^" CD8⁺ T

cells present direct, ex vivo cytolytic activity (Azuma, et al., J. Immunol. 150: 1147-1159 (1993),

and Hamann, et al., J. Exp. Med. 186: 1407-1418 (1997), have proposed that such cells

correspond to effector type CTLs. In view of these prior observations, individuals who presented

significant amounts of circulating cells positive for tetramers of SEQ ID NO: 3 or SEQ ID NO:

5 were phenotyped for CD28, CD45RA, and CD45RO.

To do this, CD8⁺ lymphocyte populations were purified, as described supra. The highly

purified samples were divided into two, and then contacted with one of the tetramers described

supra, and with antibodies against CD45RA labelled with Cy-Chrome, and either antibody

against CD45RO labelled with FITC, or antibody against CD28 labelled with FITC. The

procedures for carrying out the staining are described, supra, and are not repeated here.

Virtually all circulating CD8⁺ T cells from healthy donors were positive for CD28, and

CD45RA, and negative for CD45RO when the cells were positive for tetramers of SEQ ID NO:

3. In contrast, CD8⁺ T cells which were positive for SEQ ID NO: 5 tetramers were CD28 and

CD45RO positive, and CD45RA negative (i.e., CD45RA^low). This suggests that the CD8" T cells

specific for SEQ ID NO: 5 were antigen experienced memory cells, which is compatible with the idea that SEQ ID NO: 5 represents a recall antigen. The phenotype of the CD8⁺ T cells from

healthy patients, which are specific to SEQ ID NO: 3, in contrast, suggests a naive phenotype.

This analysis was extended to melanoma patients. The majority of circulating CD8^* T

cells specific for SEQ ID NO: 3 containing tetramers were CD28 and CD45RA positive, but

CD45RO negative. The results are presented graphically in figure 6.

Of the ten melanoma patients tested, the circulating CD8⁺ T cells specific for SEQ ID

NO: 3 in seven presented a uniformly naive phenotype (i.e., CD28 and CD45RA positive,

CD45RO negative), a phenotype identical to that of healthy donors' CD8⁺ T cells. Three of

these melanoma patients either displayed greater than 35% memory cells (CD28 and CD45RO

positive, CD45RA negative), or more than 90% effector-like cells (CD28 negative, CD45RA

intermediate levels, CD45RO negative).

Phenotype and frequency of CD8⁺ T cells positive for Melan-A tetramers were not

correlated, in general. For example, the memory cells found in the two melanoma patients were

not found at increased frequencies, while the high frequencies of CD8⁺ T cells positive for the

Melan-A containing tetramers were not of memory phenotype.

When the CD8⁺ T cells which were positive for tetramers containing SEQ ID NO: 5 were

analyzed, the memory phenotype discussed supra was found in 11 of the 13 melanoma and

melanoma/vitiligo patients tested, as well as in all healthy donors. There was some heterogeneity

in two of the melanoma patients.

Example 36

In these experiments, the staining methodology, limiting dilution assays, and the

ELISPOT assay technique, all of which are described, supra, were compared. To start, sorted cells from two healthy donors (HD 329 and HD 604, in the table which

follows), and two melanoma patients (LAU 132 and LAU 203), were compared, using the

functional assays described supra.

In the two healthy patients and one of the melanoma patients, nearly all cells were

CD45RA^'

(95, 95 and 94%), while only 54% of the second melanoma patient's cells were CD45RA⁺.

When samples were tested via limiting dilution, almost all of the Melan-A specific CTL

activity was found in the CD45RA" naive subsets (98, 99 and 90%), while 73% of the CTL

activity for patient LAU 132 was found in the CD45RA^" (CD45RA^low) subset. Distribution of

CTL precursors among naive and memory subsets, as analyzed by LDA, paralleled the results

obtained using the flow cytometry assays; however, frequency of the CTL precursors was

underestimated by LDA, especially in the CD45RA⁺ population. The level of underestimation

was by a factor of 3.6 for CD45RA^" (CD45RA^,0W) cells, but 11.4 for CD45RA^* cells.

ELISPOT assays were then carried out on unsorted PBLs, due to an insufficiency of pure

cell populations. The number of positives for Melan-A peptide antigen was extremely low,

indicating high underestimation using the technique.

In the table which follows, the values for CD45RA - (CD45RA^low) cells for HD 329, HD

604 and LAU 203 were normalized, because frequency values were below the lower limit of

detection. Melan-A specific CD8⁺ T-cell frequen cy [10-⁵]

Tetramers LDA ELISPOT

CD8⁺ subset: CD45RA⁺ CD45RA CD45RA⁺ CD45RA total

(CD45RA^|0W) (CD45RA^l0W)

HD329 98 ± 11 7 ± 2 13 ± 0.2 0.3 ± 0.0 0.9 ± 0.9 HD604 190 ± 6 12 ± 3 9.7 ± 1.9 0.1 ± 0.0 0.0 ± 0.0 LAU 132 36 33 3.3 ± 0.4 9.1 ± 0.1 5.5 ± 5.6 LAU 203 100 4 13 ± 2.9 1.3 ± 0.1 0.0 ± 0.0

Example 37

Additional experiments were carried out to provide a better comparison of the sensitivity

of the ELISPOT technique to tetramer staining. In these experiments, PBLs from the 10 healthy

donors, and twelve of the thirteen melanoma patients were analyzed, using SEQ ID NOS: 3 and

5, staining assays and the ELISPOT assay.

Significant levels of spots specific for SEQ ID NO: 5 were found in all cases (over three

times background), with deduced frequency correlating to the tetramer calculations. The

ELISPOT assay did generally underestimate the frequency of CTL precursors specific for SEQ

ID NO: 5.

On the other hand, SEQ ID NO: 3 specific spots were only detectable in one melanoma

patient. The apparent frequency of SEQ ID NO: 3 specific, ELISPOT positive cells was 93 times

lower than that obtained using tetramer staining. When some portion of the SEQ ID NO: 3

tetramer positive cells presented antigen experienced phenotype, frequency was less

underestimated (12, and 4 fold, in two assays). This appears to show that efficient detection by

ELISPOT is restricted to memory phenotype, antigen specific cells. Example 38

The fate of SEQ ID NO: 3 specific T cells in vivo over time was studied. In these

experiments, blood samples were taken from a melanoma patient over a period of two years. To

elaborate, the patient had presented primary skin melanoma of the lower limb. Inguinal lymph

node dissection revealed that 4 of 6 lymph nodes were infiltrated by melanoma cells. The patient

received isolated limb perfusion therapy (melphalan), and then adjuvant IFN alpha therapy, over

1.5 years. A second inguinal lymph node dissection showed that 15/16 lymph nodes were

infiltrated with melanoma. The patient was tumor free for nearly two years. Over that period,

he received immunizations of melanoma specific peptides (3-4 weekly in subcutaneous inj ections

of lOOμg of each of various peptides in PBS). The peptides corresponded to amino acids 26-35

of Melan-A (i.e., SEQ ID NO: 2), amino acids 1-9 of tyrosinase (i.e., MLLAVLYCLL, SEQ ID

NO: 49), amino acids 368-378 of tyrosinase (i.e., SEQ ID NO: 4), amino acids 280-288 of gp

100, (YLLEPGPVTA, SEQ ID NO: 50) and amino acids 457-466 of tyrosinase (i.e.,

LLDGTATLRL; SEQ ID NO: 51). Five cycles of immunization were given, four of which

included daily, subcutaneous administration of GM-CSF, starting 4 days before immunization,

and covering the entire period of immunization.

Prior to the first immunization cycle, SEQ ID NO: 3 tetramer positive CD8^* T cells of

the patient (0.04% of total CD8⁺ T cells), presented a naive, CD45RA" phenotype (The assay

used is as described supra, using the same purification methods, anti-CD45RA Cy-Chrome

labelled mAbs, and anti-CD28 FITC labelled mAbs).

One month after the end of the first injection, and until the end of the second cycle of

immunization, half of the tetramer" cells presented antigen experienced, CD45RA^~(CD45RA^low)

phenotype, accompanied by a small increase in frequency of SEQ ID NO: 3 positive cells (0.04% to 0.07%). During the second year, the CD45RA (CD45RA^l0W) SEQ ID NO: 3 positive cells

gradually disappeared (dropping from 51% to 23%), while the frequency of total SEQ ID NO:

3 tetramer positive cells remained constant. The vast majority of these cells continually

displayed CD28 positive phenotypes.

Example 39

The experiments described herein were designed to obtain tumor antigen specific T cells

from tumor infiltrated lymph nodes ("TILN" hereafter). Single cell suspensions were prepared

from a TILN sample taken from a melanoma patient, in accordance with Romero, et al., J.

Immunol. 159: 2366-2374 (1997), incorporated by reference. The patient had been typed,

previously, as HLA-A^T 0201 positive. The single cells were cultured in medium (2 ml of

Iscove's Dulbecco medium, supplemented with Asn, Arg and Gin, 10%> pooled, human A^* serum

and recombinant IL-2 (100 U/ml) and recombinant IL-7 (lOng/ml). After 16 days of culture, the

cells were analyzed as described supra, using tetramers containing SEQ ID NO: 3, mAbs to CD3

labelled with peridinin chlorophyll protein (PerCP), and mAbs to CD8, labelled with FITC. The

TILN were found to contain more than 90% CD37CD8⁺ cells. Of these, 19.4% were positive

for the tetramers, which is a high percentage.

Example 40

These experiments detail the expansion of the positive population described supra. First,

cells were sterile sorted into positive cells, using standard methods. The positive and negative

cells were then cultured for 2 weeks in the presence of irradiated, allogeneic PBMCs and

phytohemagglutinin, as described supra. Following the culture, the cells were retested. using tetramers and the CD8 specific mAb described supra. The positive portion contained 94.4%

tetramer positive cells, and the negative portion, 1.1% positives.

Example 41

Following the experiments described supra, the expanded populations of positive and

negative cells, as well as TILN from LAU 203 were tested in a ⁵¹Cr release assay, as described

supra, together with the peptide of SEQ ID NO:2, which was added at lμM, or without it, using

T2 cells labelled with the ⁵¹Cr as a target.

No cytotoxicity was observed without the peptide. The tetramer positive population

showed enhanced antigen specific cytolytic activity as compared to unfractionated TILN, which

was proportional to the level of enrichment. A low level of cytolysis was detected in the negative

cells, possibly resulting from a small percentage of tetramer positive T cells in the culture.

Example 42

These experiments were designed to compare the fine specificities of the unsorted, TILN

population and the tetramer positive cells described supra. This was done by carrying out ⁵¹Cr

release assays, as described supra, using T2 cells which had been incubated with varying

concentrations of different peptides, in 50 μl samples. The peptides tested were SEQ ID NOS:

1, 2, 3, 37 and 38. The T2 cells were incubated with the peptides for 15 minutes at room

temperature before adding either unsorted TILN population cells or the sorted, tetramer positive

cells, at an effector/target ratio of 10/1. Chromium release was measured after 4 hours at 37°C.

The results are presented in figures 7 A and 7B , where figure 7 A shows the results obtained using

the unsorted TILN population and 7B the sorted population. The two populations showed similar lysis patterns. Parental decapeptide (SEQ ID NO: 1) was recognized about 5 fold more

efficiently than the parental nonapeptide (SEQ ID NO: 1), but decapeptide SEQ ID NO: 3 was

recognized much more efficiently, with half maximal lysis being obtained at low peptide

concentrations, i.e., 10^"10 to 10^"" M.

These experiments show nearly identical fine specificity in sorted and unsorted TILN,

showing that the tetramer positive, sorted cell population can be considered a polyclonal,

monospecific T cells population, representative of the total T cell population which is specific

for the antigen/MHC complex of interest.

Example 43

T cell receptor down regulation is one of the earliest T cell activation events induced upon

antigen recognition and subsequent TCR triggering. The tetramer sorted population discussed

supra was used to study this.

Specifically, T2 cells were pulsed for 1 hour at 37°C with varying concentrations of

peptides, and were then washed, twice, to avoid peptide autopresentation by T cells. Samples

(1 x 10⁵) of the T cells were then stimulated by peptide pulsed T2 cells (2 x 10⁵). After ό hours

of incubation at 37°C, the cells were stained for CD3 using a fluorescent labelled monoclonal

antibody, and were then analyzed, via standard methods. The parental Melan-A peptides (SEQ

ID NOS: 1 and 2) induced only weak TCR down regulation even at the high peptide doses that

were efficient in target CTL sensitization; however, maximal TCR down regulation was induced

by SEQ ID NO: 3. Example 44

These experiments describe studies of Ca^2* responses. These responses occur a few

seconds after T cell receptors are engaged. Essentially, when T cells are activated, calcium enters

the cells, and can be measured. T cells were loaded with a commercially available fluorescent

Ca^2* indicator, Indo-1 AM, described by Grynkiewicz, et al. J. Biol. Chem. 260(6): 3440-3450

(1985) following Valitutti, et al., J. Exp. Med. 183: 1917-1921 (1996), both of which are

incorporated by reference. As a positive control, a CTL clone was used which had been derived

from a normal HLA-A2 positive donor, PBMCs of whom had been stimulated, in vitro, with

SEQ ID NO: 2. Samples of this CTL were combined with T2 cells that had been pulsed, for 2

hours at 37°C, with serial dilutions of SEQ ID NO: 2 or SEQ ID NO: 3. Cells were centrifuged

for one minute at 1500 x g, incubated for 1 minute at 37°C, and then resuspended and analyzed

by flow cytometry.

SEQ ID NO : 3 triggered a full Ca²⁺ response, but the CTL clone referred to supra required

a concentration of about 1 ng/ml. In contrast, the polyclonal, tetramer positive cells discussed

supra were triggered for Ca²⁺ at a broader range of peptide analogue concentrations. Some cells

even reacted at the lowest peptide concentration, i.e., 10^"3 ng/ml. Further, at the maximal peptide

concentrations tested, SEQ ID NO: 2 triggered Ca²⁺ influx in only 60% of the cells of the CTL

clone, and 90% of the polyclonal population. In contrast, SEQ ID NO: 3 mobilized Ca^2* in

nearly 100% of the CTL sample.

The data also showed that almost 50% of TCR down-regulation was required to attain

a full Ca^2τ response. SEQ ID NO: 3 also induced a more rapid Ca^2* response as compared to the

parental peptide. The observation that 90% of the tetramer positive population increased Ca^2" in response

to the parental decapeptide confirmed that the tetramer positive population was highly specific

for the parental peptide.

Example 45

In addition to the early activation events described in examples 43 and 44, late activation

events were also studied, i.e., cytokine synthesis, and induction of cytolytic effector cell function.

The former was done via intracellular staining of cytokines, which is described by Jung, et al,

J. Immunol. Meth. 159:197-207 (1993), as a very sensitive method for determining cytokine

response at the single cell level. Labelled monoclonal antibodies against TNF-alpha, INF

gamma, GM-CSF, and IL-4 were used. All mAbs were labelled with phycoerythrin. In brief,

5 x 10⁴ T cells were stained with different mAbs, at concentrations of 5mg/ml, for 30 minutes

at room temperature. Dilutions and washes were carried out using PBS 10.1 %, containing 0.1%

saponin. The saponinpermeabilizes the cell, allowing entry of the mAbs. The production of each

of these cytokines by the CTL line described supra, and the polyclonal, tetramer positive

population, in response to different peptide concentrations were measured. Peptides defined by

SEQ ID NOS: 1, 2 and 3 were tested. The results are presented in figure 8. "Clone 1.13" is the

CTL clone referred to supra.

None of the peptides stimulated the entire monoclonal or polyclonal population, which

is consistent with prior reports that a fraction of CTLs is refractory to cytokine production. SEQ

ID NOS: 1 and 2 only stimulated small fractions of the CTLs (e.g., only 30%_o of the cells of the

CTL clone produced TNF-alpha at the highest concentrations of these peptides). Similarly, 45%

and 10% of the polyclonal population produced TNF-alpha in response to SEQ ID NO: 1 and SEQ ID NO: 2, respectively. In contrast, SEQ ID NO: 3 stimulated the totality of CTLs able to

produce cytokines, with a plateau being reached at 1 μM, for each population. It is interesting

to note that 45%o of the polyclonal population produced some TNF-alpha, while 90% of these

cells increased intracellular Ca^2* concentration at the highest concentration of SEQ ID NOS: 1

and 2 used.

The percentage of the population which produced IL-2 and IL-4 was always less than the

percentage producing TNF-alpha, regardless of the peptide concentrations used.

The polyclonal population was distinguishable from the monoclonal population in that

it had a greater capacity to produce IFN-gamma, and was incapable of producing IL-4, which is

consistent with prior observations.

Double staining experiments were carried out on both populations, following stimulation

with 10 μM of SEQ ID NOS: 1, 2, and 3. The double stainings were IFN-gamma and TNF-

alpha, and IFN-gamma and IL-2. SEQ ID NOS: 1 and 2 stimulated low fractions of these cells

to produce the cytokines; however, when SEQ ID NO: 3 was used, both the percentages of cells

producing the cytokines, and the amount of cytokine produced per cell were increased.

Most cells produced TNF-alpha and IFN-gamma, while only a fraction of IFN-gamma

positive cells produced IL-2. More importantly, however, the fraction of IL-2 producing cells

and the amount of IL-2 produced on a per cell basis were highly increased after stimulation with

SEQ ID NO: 3.

Example 46

It is known that full CTL activation is associated with cell division and induction of

cytolytic effector function. Such functions require strongly agonistic peptides, and/or higher peptide concentrations, in order to maximize induction of expansion and cytolytic effector

function, as compared to target sensitization, or induction of cytokine synthesis. In the case of

the peptides being studied herein, the low potential of SEQ ID NOS: 1 and 2 to induce cytokine

production by specific CTLs may mean that these peptides are inefficient activators of cytolytic

effector function. These experiments were designed to test this.

Peripheral blood lymphocytes of patent LAU 203 were stimulated with syngeneic

stimulator cells (PBMCs) that had been pulsed for 2 hours at 37°C in serum free medium together

with lμM of one of SEQ ID NOS: 1, 2, 3, 38 or 39, and 3 ug/ml of human β2-microglobulin.

These peptide pulsed PBMCs were washed, twice, irradiated (3000 rads), and then adjusted to

an appropriate volume before adding to the responder cells. In addition to the stimulator cells,

IL-2 (10 U/ml), and IL-7 (10 ng/ml), were added during the first two stimulation cycles (week

1 and week 2), and then IL-2 alone (100 U/ml) was added during the third week.

After 3 weeks of incubation, the PBMCs were analyzed via flow cytometry. The results

showed that cultures that had been stimulated with SEQ ID NO: 1 or 2 had 1.7% and 3.5%

tetramer positive cells, while those populations stimulated with SEQ ID NOS: 3, 38 and 39 had

11.2% (SEQ ID NO: 3), 19.7% (SEQ ID NO: 38), and 37.0% (SEQ 10 NO: 39) percent

positives.

Cytotoxicity was also tested, in ⁵¹Cr release assays as described supra. The cytotoxicity

induced by SEQ ID NOS : 1 and 2 was weak, while analogue peptides induced strong cytotoxicity

against target cells labelled with parent peptides SEQ ID NOS: 1 and 2, and autologous tumor

cells. This strong cytotoxicity was caused by high frequency of antigen specific cells in the

culture, and strong activation of these cells. Example 47

Experiments were carried out to study inhibition of CTL function. More particularly,

notwithstanding the observations described in the previous examples, it is known that in vivo T

cell activity does not protect most cancer patients.

There has been some speculation that cytolytic function is inhibited through the activity

of natural killer ("NK" hereafter) receptors. Two families of these have been observed, i.e., type

I transmembrane proteins, which belong to the imrnunoglobulin (Ig) superfamily, such as p58.2

(Moretta, et al., J. Exp. Med. 182:875-884 (1995)) and ILT2 (Colonna, et al., J. Exp. Med.

186:1809- 1818 (1997)), and those which are type II transmembrane proteins which contain a C-

type lectin domain, such as the CD94/NKG2 heterodimer. (Maretta, etal., J. Exp. Med. 180:545-

555 (1994)) . The experiments in this example were designed to determine whether NK receptors

may interfere with in vivo tumor specific responses.

First, CD8^* cells were examined for NK receptor phenotype. To do this, blood samples

and lymph nodes were obtained from patients with advanced stage malignant melanoma, using

standard techniques. (The patients had been typed as HLA-A*0201 positive via flow cytometry

of PBMCs, using mAb BB7.2, described supra.)

PBLs were separated from blood samples taken from patients who suffered from vitiligo

as well as melanoma via centrifugation over Ficoll-Paque, washed 3 times, and then

cryopreserved in RPMI 1640, 40% fetal calf serum, and 10% DMSO. Vials containing from 5-

lOxlO⁶ cells were stored in liquid nitrogen.

Lymph nodes were collected by surgical dissection, dissociated into single cell

suspensions in sterile RPMI 1640 supplemented with 10% fetal calf serum, washed and

cyropreserved, as described supra. Aliquots of cells were placed in 24 well tissue culture plates in 2ml of Iscove's Dulbecco medium, supplemented with 0.55 mM Arg, 0.24 mM Asn, 1.5 mM

Gin, 10% pooled human A^{^}serum, recombinant human IL-2 ( 100 U/ml), and recombinant human

IL-7 (lOng/ml).

Both CD8^* cells and the cells taken from the lymph nodes were contacted with

monoclonal antibodies against p58.2/CD158b, CD94, CD94/NKG2A, and ILT2, each of which

is an NK receptor, and tetramers of SEQ ID NO:3 labelled with phycoerythrin.

The TILNs contained high percentages of CD8^* tetramer ^* cells (1.3-5.1%). In one of the

samples analyzed, these were largely negative for p58.2 and ILT2, but were highly positive for

CD94 and CD94/NKG2A. A sample of TILN from another patient expressed some ILT2, but

low levels of the others. Yet a third sample expressed ILT2, CD94, and CD94/NKG2A.

The PBLs were found to contain from 0.10 to 0.17% CD8^* tetramer ^~ cells. One sample

was negative for all of the receptors. The CD8^* tetramer ^* cells of a second sample were almost

all positive for CD94 and CD94/NKG2A, and about half expressed ILT2. A third sample of

PBLs showed a low percentage of NK receptor positive cells.

Example 48

The relatively high frequency of tumor specific CTLs in the vitiligo patients permitted

investigation of the in vivo phenotype of these cells. Fluorochrome labelled anti-CD45RA mAbs

were used, in an assay as described supra. Two samples had high levels of CD45RA positive

cells, which are indicative of naive T cells, while reduced levels were found in a third.

Additional assays were carried out using fluorochrome labelled mAbs against CD28 and

adhesion molecule CD57, since downregulation of CD28 and upregulation of CD57 have been

recognized in activated effector CTLs. It was found that samples taken from the one vitiligo subject, described supra, with a high

percentage of NK receptor expressing CTLs were predominantly CD28^" and CD57⁺, while the

other two subjects were CD28^* and CD57^".

Example 49

Given that the NK receptors had been typed, it was important to determine if they were

functional. Previously, it had been shown that antigen specific, T cell cytotoxicity can be

inhibited through the triggering of the receptor CD94/NKG2A (Noppen, et al., Eur. J. Immunol.

28:1134-1142 (1998)); or p58.2 (Phillips, et al., Science 268: 403-405 (1995); Ikeda, et al,

Immunity 6: 199-208 (1997), Bakker, et al., J. Immunol. 160:5239-5245 (1998)). Experiments

were developed, and are described herein, to investigate cytotoxicity of cells ex vivo, directly

after withdrawal from the patient. PBLs were sorted from two different samples, and sorted via

FACS, using methodologies described supra, using anti CD8 and either anti p58.2 or anti

CD94/NKG 2A antibodies. The sorted suspensions were cultured, overnight, in Iscove

Dulbecco's medium supplemented with 0.55 mM Arg, 0.24 mM Asn, 1.5mM Gin, 10% pooled

human A^* serum, and recombinant human IL-2 (10 U/ml).

Cytolytic activity was then tested in anti-CD3 mAb redirected ⁵¹Cr release assays. In

these assays, Fcγ-receptor expressing P815 mastocytoma cells were radiolabelled with Na⁵ ' CrO₄

for 1 hour at 37°C, at 5% CO₂. After washing, 10³ P815 target cells were coincubated with PBLs

which were either CD94/NKG2 A positive or CD94/NKG2 A negative, at varying effector: target

levels, together with varying concentrations of anti-CD 3 mAbs. The anti-CD3 mAbs cross link

CD3 molecules on T cells to the Fcγ-receptors, leading to the lysis. Samples were incubated for 4 hours at 37°C, after which supernatant were collected and radioactivity measured. To

determine the role of the receptor, anti CD94/NKG2A antibody was either added, or withheld.

As a control, anti-CD 19 antibodies were used.

The results are shown in figure 9. Squares represent tests carried out using anti

CD94/NKG2A antibody, and triangles, the control, anti-CD 19 antibody.

In the presence of the anti-CD94/NKG2A antibody, the cells positive for this receptor

showed reduced killing, as compared to killing in the presence of anti CD 19 antibody, indicating

that the antibody stimulated the NK receptor, i.e., the CD94/NKG2A receptor. In contrast,

cytolytic activity of those CD8^ cells negative for this receptor were not reduced in the presence

of the antibody. Further, the NK receptor positive population showed stronger cytolytic activity,

as compared to the NK receptor negative population.

Example 50

Phenotypic analysis of various samples revealed one which had an unusually high

frequency of αβTCR^*, CD28 negative cells which expressed p58.2. These cells provided a

singular opportunity to test functional consequences of p58.2 binding to its natural ligand HLA-

Cw3. As in example 49, cells of this sample were sorted via FACS, into CD8^* p58.2^* and CD8^*

p58.2^" populations, and were tested as described supra, but without any antibodies. P815 cells

were used, as were P815 cells transfected to express HLA-Cw3, using known methods.

The P815 cells were killed efficiently by both p58.2 positive and negative cells at

effector: target ratios of 5.1 ; however, lysis of transfected cells was reduced markedly, indicating

inhibition of lysis via interaction of p58.2 with HLA-Cw3. In the absence of CD3 specific

antibodies, killing was below 2%. The foregoing examples describe aspects of the invention, which is a multicomponent

complex useful, e.g., in isolating cytolytic T cells specific for a particular target, from a sample.

The complex comprises a first binding partner and a second binding partner, wherein the first and

second binding partner are specific for each other. These can be, e.g., avidin or streptavidin and

biotin, an antibody or a binding portion of an antibody specific to biotin, and so forth. The key

feature is that the second binding partner must be bound to a plurality of complexes of an MHC

molecule, a B2 microglobulin molecule and a peptide which binds specifically to said MHC

molecule, and the multicomponent complex must be labelled. The MHC molecules are

preferably HLA molecules, such as HLA-A2 molecules. The examples all show HLA-A^*0201,

however, it will be understood by the artisan of ordinary skill that any HLA molecule could be

used. With respect to the peptide of interest, many references, including review articles, U.S. and

non-U.S. patents, and so forth describe peptides beyond SEQ ID NOS: 2-5 and their binding

partner HLA molecule. All are encompassed by the invention. Exemplary peptides and their

HLA molecule partners are presented later in this application.

Preferably, the second binding partner is biotin, but it may also be, e.g., an antibody

which is specific for a component of the HLA B2 microglobulin peptide complex, such as an

HLA specific antibody, or a B2 microglobulin specific antibody. Similarly, the first binding

partner may be e.g., recombinant or naturally occurring protein L, recombinant or naturally

occurring protein A, or even a second antibody. The complex can be in soluble form, or bound,

e.g., to a removable solid phase, such as a magnetic bead.

The number of HLA/B2 microglobulin/peptide complexes in the large molecule of the

invention may vary. It comprises at least two complexes, and preferably at least four, but more

may be present as well. The complex of binding partners and HLA/B2 microglobulin/peptide may be labelled,

using any of the labels known to the art. Examples of fluorescent labels are given supra.

Enzymatic labels, such as alkaline phosphatase, metal particles, colored plastics made of

synthetic materials, radioactive labels, etc., may all be used.

A third binding partner may also be used, which binds, specifically, to the first binding

partner. For example, if the first binding partner is streptavidin, and the second binding partner

is biotin, then the third binding partner may be a streptavidin specific antibody. When three or

more binding partners are used, the label refeπed to supra may be attached to any of the binding

partners so long as engagement with the HLA/B2 microglobulin peptide complexes is not

impaired.

The complexes may be used, e.g., to identify or to isolate cytolytic T cells present in a

sample, where these cells are specific for the HLA/B2 microglobulin/peptide complex. As the

examples show, such cytolytic T cells bind to the immunocomplexes of the invention. In a

preferred embodiment, the sample being tested is treated with a reactant which specifically binds

to a cytolytic T cell, wherein said label provides a detectable signal. The sample, including

labelled CTLs, is then contacted to the complex, where it binds, and can be separated via any of

the standard, well known approaches to cell separation. Preferably, FACS is used, but other

separation methodologies will be known to the skilled artisan as well. The peptide used is left

to the needs of the skilled artisan, and will depend, e.g., on the nature of the specific MHC

system under consideration, a table of exemplary, but no means the only, peptides for which

CTLs are known, follows. These are also set forth as SEQ ID NOS: 6-36. Gene MHC Peptide SEQ ID

MAGE-1 HLA-A1 EADPTGHSY 6

HLA-Cwl6 SAYGEPRKL 7

MAGE-3 HLA-A1 EVDPIGHLY 8

HLA-A2 FLWGPRALV 9

HLA-B44 MEVDPIGHLY 10

BAGE HLA-Cwl6 AARAVFLAL 11

GAGE- 1,2 HLA-Cwl6 YRPRPRRY 12

RAGE HLA-B7 SPSSNRIRNT 13

GntV HLA-A2 VLPDVFIRC(V) 14

MUM-1 HLA-B44 EEKLIVVLF 15

EEKLSVVLF 16

CDK4 HLA-A2 ACDPHSGHFV 17

ARDPHSGHFV 18

B-catenin HLA-A24 SYLDSGIHF 19

SYLDSGIHS 20

Tyrosinase HLA-A2 MLLAVLYCL 21

HLA-A2 YMNGTMSQV 22

HLA-A24 AFLPWHRLF 23

HLA-B44 SEIWRDIDF 24

HLA-B44 YEIWRDIDG 25

HLA-DR4 QNILLSNAPLGPGFP 26

HLA-DR4 DYSYLQDSDPDSFQD 27

Melan-A^Mart-' HLA-A2 (E)AAGIGILTV 28

HLA-A2 ILTVILGVL 29 gpl00^Pmel117 HLA-A2 KTWGQYWQV 30

HLA-A2 ITDQVPFSV 31

HLA-A2 YLEPGPVTA 32

HLA-A2 LLDGTATLRL 33 HLA-A2 VLYRYGSFSV 34

DAGE HLA-A24 LYVDSLFFL 35

MAGE-6 HLA-Cwl6 KISGGPRISYPL 36

Additional peptides may be found, e.g., in U.S. patent application Serial Nos. 08/672,351,

08/718,964, 08/530,569, and 08/880,963, all of which are incorporated by reference, as well as

U.S. Patent No. 5,821,122, also incorporated by reference.

Additionally, the method can be used to monitor the status of tumors, following

administration of a particular therapeutic agent, such as a vaccine. Further, since the

methodology can be used to identify cytolytic T cell precursors, as shown, supra, one can identify

candidates for potential therapies by determining if they possess the relevant T cell precursors.

Also apart of the invention is the use of tetramers as described, in conjunction with other

steps, to yield populations of T cells with desired features. These include specific phenotypes,

such as phenotypes associated with antigen experienced, or memory cells, or naive cells, and so

forth. Such populations can be cultured, in the absence of peptides, to yield cell populations

which can be used, e.g., diagnostically or therapeutically. This culturing can be carried out, e.g.,

with a mitogen, such as phytohemagluttinin, and without peptides.

The invention also involves methods for obtaining desired T cells via in vivo recruitment.

The examples show, e.g., that one can inject a peptide of interest into a subject, e.g.,

subcutaneously, leading to recruitment of T cells. In turn, the T cells can be separated, and

cultured, in the same manner described supra.

The invention also involves the ability to improve the lytic activity of a T cell population.

As has been shown, one can assay a population of T cells, to determine if these T cells express

NK receptors. Expression of said receptors allows one to gauge the lytic potential of the T cell population involved. An antibody assay is one way to do this, but hybridization assays, and other

approaches may also be used.

Once the NK receptor expression levels are determined, one can improve the lytic ability

of the population by adding NK receptor inhibitors, such as soluble, MHC molecules like soluble

HLA-A* 0201 or HLA-Cw*3 molecules, or agents which permit the T cell to function

notwithstanding expression of NK receptors. An example of such an agent, as was shown, is an

anti-CD3 antibody.

Other aspects of the invention will be clear to the skilled artisan, and need not be

elaborated further.

The terms and expressions which have been employed are used as terms of description

and not of limitation, and there is no intention in the use of such terms and expressions of

excluding any equivalents of the features shown and described or portions thereof, it being

recognized that various modifications are possible within the scope of the invention.

Claims

Claims

1. An isolated complex comprising a first binding partner and a second binding partner

which bind to each other, at least one of which is labelled to provide a detectable signal,

and a plurality of immune complexes comprising an MHC molecule, a B2 microglobulin

molecule, and a peptide which specifically binds with said MHC molecule.

2. The isolated complex of claim 1 , wherein said second binding partner is biotin.

3. The isolated complex of claim 1, wherein said first binding partner is avidin or

streptavidin.

4. The isolated complex of claim 1, comprising at least four of said immune complexes.

5. The isolated complex of claim 4, wherein said immune complexes are identical to each

other.

6. The isolated complex of claim 1, wherein said complex is labelled with a fluorescent or

enzymatic label.

7. The isolated complex of claim 1, wherein said immune complex is a complex found on

surfaces of cancer cells.

8. The isolated complex of claim 1, wherein said first binding partner is an antibody or a

binding fragment of an antibody.

9. The isolated complex of claim 1, wherein said second binding partner is an antibody.

10. The isolated complex of claim 1, further comprising a third binding partner which

specifically binds to said first binding partner.

11. The isolated complex of claim 10, wherein said third binding partner is an antibody.

12. The isolated complex of claim 1 , further comprising a solid phase to which said isolated

complex is bound.

13. The isolated complex of claim 1 , wherein said solid phase is a magnetic bead.

14. A method for determining presence of cytolytic T cells specific for an immune complex

of an HLA molecule, a B2 microglobulin molecule and a peptide, comprising contacting

a T cell containing sample with the isolated complex of claim 1 , and determining binding

of cytolytic T cells to said isolated complex as a determination of said cytolytic T cells

in said sample.

15. The method of claim 14, further comprising labelling said cytolytic T cells with a

detectable label which specifically binds to cytolytic T cells.

16. The method of claim 15 , wherein said label which specifically binds to cytolytic T cells

is an anti-CD8, anti-CD28, anti-CD45RO, or antiCD45RA antibody.

17. A method for isolating cytolytic T cells which specifically bind to a complex of an HLA

molecule, a B2 microglobulin molecule, and a peptide, comprising contacting a sample

containing said cytolytic T cells with the isolated complex of claim 1 to bind cytolytic

T cells thereto, separating said isolated complex from said sample, and removing any

cytolytic T cells bound thereto.

18. A method for monitoring status of a tumor, comprising contacting a T cell containing

sample from a subject with a tumor with the multimeric complex of claim 1 at a first

point in time, repeating contact at a second point in time, and determining variation in

T cell type or number at said second point in time as compared to T cell type or number

at said first point in time as a determination of tumor status.

19. The method of claim 18, further comprising administering a therapeutic agent to said

subject after said first point in time and before said second point in time.

20. A method for identifying a candidate for treatment of a condition characterized by a T

cell response to said condition, comprising contacting a T cell containing sample from

said candidate with the multimeric complex of claim 1, determining any cytolytic T cell

precursors which bind thereto, and determining whether any of said cytolytic T cell

precursors are of a type sufficient to alleviate said condition.

21. A method for preparing a population of cytolytic T cells specific for a complex of an

MHC molecule and a peptide, comprising:

(a) contacting a sample containing T cells specific for said complex with the

complex of claim 1, under conditions favoring binding of said T cells to said complex,

(b) removing said T cells, and

(c) culturing said T cells in a medium which contains no peptides.

22. The method of claim 21 , wherein said complex is a tetramer.

23. The method of claim 22, wherein said tetramer comprises four MHC molecules,

and four peptide molecules, each of which is complexed with one of said four

MHC molecules.

24. The method of claim 23, wherein said four peptide molecules are tumor rejection

antigens.

25. The method of claim 21 , further comprising adding an amount of said peptide to said

sample prior to contacting with said tetramer, for a time sufficient to stimulate T cells in

said sample.

26. The method of claim 21 , wherein said medium contains a mitogen.

27. The method of claim 26, wherein said mitogen is phytohemaglutinin.

28. An isolated peptide, the amino acid sequence of which consists of an amino acid

sequence set forth in one of SEQ ID NOS: 39-47.

29. The isolated peptide of claim 28, the amino acid sequence of which is set forth at SEQ

ID NO: 41, 43, 45 or 46.

30. The isolated complex of claim 4, comprising 4 molecules of SEQ ID NO: 1, 2 , 3 or 4,

each of which is complexed to an HLA-A* 0201 molecule.

31. The method of claim 25, wherein said peptide is the peptide of SEQ ID NO: 4.

32. The isolated complex of claim 4, comprising 4 molecules of SEQ ID NO: 9, each of

which is complexed to an HLA-A* 0201 molecule.

33. The method of claim 25, wherein said peptide is the peptide of SEQ ID NO: 9.

34. A method for obtaining a population of T cells, comprising, injecting skin of a subject

with an amount of a peptide sufficient to stimulate recruitment of cytolytic T cells to the

site of injection, taking a skin biopsy from said site of injection, separating any

lymphocytes from said biopsy, isolating a subset of lymphocytes specific for complexes

of said peptide and an MHC molecule from said lymphocytes, by contacting the

lymphocytes with the isolated complex of claim 4, and stimulating said subset of

lymphocytes.

35. The method of claim 34. wherein said subject is afflicted with cancer.

36. The method of claim 34, further comprising separating said subset of lymphocytes into

naive cells and memory cells.

37. The method of claim 34, comprising stimulating said subset of lymphocytes with an

amount of the peptide used for injection.

38. The method of claim 34, comprising stimulating said subset of lymphocytes in the

absence of any peptide and in the presence of a mitogen.

39. The method of claim 34, further comprising separating said subset of lymphocytes by

determining cell surface markers thereon.

40. A method for stimulating production of T cells of interest, comprising contacting a cell

sample believed to contain T cells with the isolated complex of claim 4, separating any

T cells which bind to said complex from said samples, phenotyping said T cells based

upon cell surface markers, separating a population of said T cells which possess cell

surface markers of interest, and stimulating proliferation of said population.

41. The method of claim 40, wherein said cell surface molecules of interest are markers

which indicate that said T cells are antigen experienced.

42. A method for assessing lytic ability of a T cell population, comprising assaying said T

cell population to determine NK receptor expression by said T cell population.

43. The method of claim 42, comprising assaying said T cell population with an antibody

which specifically binds to an NK receptor.

44. The method of claim 43 , wherein said NK receptor is p58.2 or CD94/NKG2 A.

45. A method for improving lytic activity of a T cell population, comprising adding an agent

to said T cell population which inhibits activity of any NK receptors expressed by said

T cell population.

46. The method of claim 45, wherein said agent is a soluble MHC molecule or a portion of

an MHC molecule which binds to said NK receptor.

47. The method of claim 46, wherein said agent is a soluble HLA-A2 or HLA-Cw*3

molecule.

48. A method for improving lytic activity of a T cell population, at least a portion of which

express NK receptors, comprising adding an agent to said T cell population which

facilitates binding of a T cell to its target.

49. The method of claim 48, wherein said agent is a cross-linking antibody.

0. The method of claim 49, wherein said cross-linking antibody is an anti-CD3 antibody.