CA2180287A1 - T-cell epitopes - Google Patents

Abstract

T-cell epitopes are determined by assaying pools of peptides derived from the full length antigen sequence in the absence of xenogeneic serum, and determining the mitogenic effect of the peptides.

Description

Wo 95/18148 218 0 2 8 7 PCT/US93/11703 T-Cell F~pitopes r~P~rr~tinn TPrhnir~l FjP,lA
This invention relates to the fields of molecular bio~ogy and -' "y. More specifically, this invention relates to methods for ~' g T-cell epitopes and the specific peptide epitQpeS (' ' thereby.
R~.-k~rnl~n~l Qf the Ir ~
T-cell ~ ....h~ have for the most part been previously studied using clones of T-cells. By definition, the cells of such a clone are I ~,, and by their very nature are not ~ f,..~live of T-cells in the general population. Thus, ~nPr~li7sltion of discoveries made with these clones to the population as a whole is not pQssible.
A knowledge of the epitopes that stimulate T-cells in the general population would be invaluable in designing new diagnostic and therapeutic agents. For example, detection of an abnormally high proportion of T-cells to a particular antigen would allow earlier diagnosis of exposure to the antigen than standard tests requiring the production of specific antibodies.
Similarly, detection of a population of T-cells sensitive to an antigen also would indicate a previous exposure, even when any antibody produced had become ~ f ~f' ~ Such methods would also be useful for monitoring the progress of v~ ~nowledge of the epitopes that stimulate killer T-cells against particular antigens would create a new class of very specific anti-c mcer agents. However, in order to use such agents, the T-cell .1..~ ....i" ~ for each individual must be identified before treatment may commence.
I

Wo 95118148 ~ PCT/US93111703 218~87~
One r ~ problem in ~' l" T-cell epitopes in the general population is obtaining sufficient ly, ' - y~ to carry out systematic studies. r llh~l~llolc, in every sample there will be an unknown and l ' ..~lc number of antigen-specific memory T-cells. In addition, there will be responses of some cell p. ,~ at a low frequency to 5 ~ , of the medium added to support cell growth in the assay. Thus, each test must be repeated a number of times to ensure that a reliable conclusion can be drawn. In many cases the supply of ly, ' yt~.~ will be lunited. For inst~Lnce, in clinical conditions (for instance, infection with HIV) the patient may be able to prQvide samples containing only a small number of reactive T ce~ls.
F. Sinigaglia et al., Meth Enzymol (1991) 203:370-86 disclosed methods for , T cell epitopes. D. Valmori et al., J Immunol (1992) 149:717 21 disclosed T
cell epitopes for tetanus toxm. A. Kumar et al., J ImmunQI (1992) 148:1499-505 disclosed "universal" T cell epitopes to gpl95 frQm P. falciparum (merQzoite surface antigen).
15 Brief Description of the Fi~ures Fig. I depicts the sequence of HIVs,~2 gpl20 (31-509), used in Example 3.
Disclosure of the Invention One aspect of the invention is a method for identifying T cell fl. ~.. ;.. - .l~, whic~
20 method comprises reacting each of a plurality of pools with an -~O l agent, th~
agent being -"O "y reactive with the antigen, each pool comprising a plurality of Ov~,lld~ g peptide sequences of the antigen, assessing the strength of reaction between tlle agent and each of the pools, selecting one or more pools giving the strQngeSt reaction with the agent, preparing a plurdlity of sub-pools, each sub-pool comprising one or a plurality of 25 peptide sequences selected from one of the selected pools, reacting cach of the sub-pools with the agent, and assessing the strength of reaction between the agent and each of the sub-pools. The agent is preferably a population or sample of peripheral blood ~' cells (PMBCs), where the strength of reaction is gauged by ~' ~ the amount of T cell activation that results.

wo 9~/1814~ 21 8 Q 2 8 7 PCTIUS93/11703 Another aspect of the invention is a method for detecting the exposure of a subject to an antigen or pathogerl by .' ~ ~ the response of a T cell-containing sample. Another aspect of the invention is a method for detecting the exposure of a subject to an antigen or pathogen by detectirlg bimding of a T cell epitope peptide to a T cell antigen receptor.
S Another aspect of the invention is an assay kit for detecting the exposure of a subject to an antigen or pathogen, comprising a T cell epitope, preferably im ~ with means for detection of a mitogerlic response or surface binding. Another aspect of theinvention is an improved vaccime . , which comprises a T cell epitope peptide inC, ~ with B cell epitope peptide.
Another aspect of the invention is a method for inducing immunity in a bird or mammal, by -' ~ a ~ùlllpu~i~iull comprising a T cell epitope and a B cell epitope specific for a pathogen.
Another aspect of the invention is a method to increase the number of T cells capable of responding to a pathogen e~C vivo, by contactimg T cells obtained from a subject with a T
cell ,' t, culturing the reactive T cells to imcrease the number of T cells capable of responding to the antigen, and Al~ t l~ said T cells to the subject.
Mn~lPcgf ~'Aryint Out Tl1P InVPnt;f~n A. D~f nitir,nc The terms "T cell epitope" and "T cell 1' - " refer to peptides or regions within a longer protein which bind T cell antigen receptors in ~ ~. with liAn MHC proteins. Preferably, T cell epitopes are l,IIAI-~ of a pathogen or malignancy.
A "T cell epitope peptide" is a peptide of about 6 to about 20 amino acids, preferably about 8 to about 15 amino acids, which primarily consists of a T cell ~ T cell epitopes are prepared from the primary sequence of antigens. A T cell epitope "having a sequence derived frgm" an antigen is a peptide which comprises a sequence of amino acids found con-secutively within the antigen's primary sequence.
The terms "B cell epitope" and "B cell ~t~ .~.." refer to antigens which are u~ ive with an antibody or B cell surface antigen receptor (membrane-bound WO 95/18148 pCTNS93/11703 2~
antibody). B cell epitopes/(' need not be proteims or peptides, but may include lipids, ~uboll~dldLriD~ and other molecules.
The term "vaccme" as used herein refers to a ~ ~ ~ or rl l..tj~Q suitable for - ' to a mammal or bird, and capable of inducing an immune response in the 5 mammal or bird. Vaccines of the invention will include at least one T ceU epitope and at least one B cell epitope, which need not be derived from the same antigen or pathogen. The B cell antigen may be presented as a whole protein or large fragment, or as whole killed pathogen, if desired. The T cell epitope is preferably included as a T cell epitope peptide.
The terms "label" and "detectable label" as used herein refer to any atom or molecule 10 which can be used to provide a detectable (preferably ~ iri-l-lP) signal, aQd which can be attached to a protein or peptide. Labels may provide signals detectable by lluu~rD~r~ ri, radioactivity, ~ lri~ly, X-ray diffraction or absorption, m:l~r~PtiC~, enzymatic activity, arld the like. Suitable labels include r~ u~ OlriS, ~ ulllu~ iD7 radioactive atoms (particularly 32p and 12sI), electron-dense reagents, enzymes, and ligands having specific 15 binding partners. Enzymes are typically detected by their activity. For example, h~-r5Pr~rlich peroxidase is usually detected by its ability to convert 3,3',5,5' t~ du~ llyil ' - (TMB) to a blue pigment, ~ with a D~ ,llul~h ~ .... t~ ., It should be understood that the above description is not meant to categorize the various labels into distinct classes, as tlle same label may serve in several different modes. For example, l2sI may serve as a radioact-20 ive label or as an electron-dense reagent. HRl? may serve as an en2yme or as an antigen for a MAb. Further, one may combine various labels for desired effect. For example, MAbs and avidin also require labels in the practice of this invention: thus, one might label a peptide with biotin, and detect its presence with avidin labeled with 12sI, or with an anti-bio-tin MAb labeled with HRP. Other r~ - and pUDD;IJ;li~ D Will be readily apparent to 2~ those of ordinary skill in the art, and are considered as equivalents within the scope of the instant invention.
The term "biological sample" refers to any sample obtained from a bird or mammalwhich contains live T cells, preferably peripheral blood ~' cells. Suitable biological samples are typically derived from blood, but may be derived fro~ ny biological ~ WOsSJ181it8 180287 pcTluss3lll7o3 ., tissue or fluid (e.g., biopsy specimens, Iymph, pus, saliva, semen, and the like) where T
- cells may be found.
A "kit" within the scope of this invention includes at least one T celi epitope peptide, and printed ~.~liu..,, for ~ , ~ am assay. "Printed il,D~Iu.,liùns'' may be written or 5 printed on paper or other media, or committed to electronic media such as magnetic tape, computer-readable disks or tape, CD-ROM, and the like. Kits also prefeMbly include means for detecting positive responses, for example labels for the T cell epitope peptides, 3H-T, or similar means. Kits may also include culture dishes, culture reagents, and other such supporting materials.
The term "I ' .... ~ 11y acceptable carrier" means refers to - , ,u.. ds and com-positions which may be ' ~i to mammals and/or birds without undue toxicity.
Exemplary ~ " ~ i. Aily acceptable saits include mineral acid saits such as hydro-chlorides, I~yJluiJlulllid~ ' , ' sulfates, and the like; and the salts of organic acids such as acetates, ~JlUr malonates, benzoates, and the like. A thorough discussion of 15 ~ ", ~ ~...li. Ally acceptable excipients is available im Remington's Pilqrrnq~ putical Sciences (Mack Pub. Co.).
The term "effective amount" refers to the amount of T cell epitope peptide and B cell antigen required to effect an immune response in a subject. The precise effective amount will vary from subject to subject, depending on age, species, si~e, weight, and general 20 health, but will generally correspûnd tû the amount effective for tMditional vaccines.
B. GPn~rAl ~P h,.,i . ~
This invention involves the uu..ll, of various test I , ' in a plurality of pools, and testing the pools for ,activity with T cells. The pools are tested for 25 activity, and selected active pools ~ y ' 1 as a plurality of sub-pools, generally derived from the sequences of the furst pool. The pluMlity of sub-pools is then assayed, and active sub-pools selected for resynthesis as a pluMlity of sub-sub-pools. This process is reiterated as many times as desired, preferably until the sub-pools contain only a single species of peptide, thus identifying the active peptides. It is essential to minimize the number of assays 30 performed, as one must use peripheMI blood Iyl..~ u~yt~ (PBLs) to obtain an accurate ... . _ . .. _ . , wo 95/18148 - PCI~/US93/117~3 ~Q2~
reading. This process minimizes tl~e number of assays which must be performed, thus making possible T cell epitope l' for an individual.
The size of the pool will depend on the number of active test ~.,.,.l..,.".~l~ expected to be discovered. Obviously, the major benefit of the invention will be realized where the 5 number of active test I . ' is a small proportion of the . . ' to be screened. To minimize the amount of test material required for the assays, the number of test comrn ~C
in each pool should be chosen so that the majority of pools will be inactive.
T cell epitopes are primarily or solely composed of peptides. Protein antigens are digested by antigen-presenting cells (APCs), and fragments of the antigens presented on the 10 APC surface in the context of a major 1~ y complex (MHC) protein. Thus, T
cell epitopes are always linear fragments of the native antigen: there are no ~ i.",.,..c epitopes, as in the case of B cell epitopes. Accordingly, T cell epitopes may be determined by preparing a series of peptides which span the amino acid sequence of the antigen. The peptides are selected to overlap, so that l' are not missed due to straddling the 15 junction between two test peptides. In general, one should allow at least 12-13 residues for a T cell ~ Thus, lf the peptides used are 13mers, one should prepare overlapping 13mers that overlap by 12 residues (i.e., that are offset by one residue, e.g., IXXXXXXXXXXXXX13, ~XXXXXXXXXXXXX14, etc.). If longer peptides are used, tlle offset may be increased accordingly. Thus, if one employs 15mers, one may synthesize a 20 series of peptides in which each is offset from the next by two or three residues (e.g., I X X X XX X X XXXXX~XXI S, 3XXXXXXXXXXXXXXX17, etc. ).
The peptides may be synthesized by any convenient method. A presently preferred method is the synthetic scheme disclosed in WO90/09395, in which peptides are ~
bound to a plurality of plastic pins via a cleavable linkage. Following synthesis, the peptides 25 are cleaved from the suppolt and are combined into pools for testing. For ease of analysis, the peptides are preferably grouped into pools on the basis of their sequences, i.e., the first ten sequential peptides in pool 1, the next ten peptides in pool 2, etc. However, other grouping strategies may also be useful. For example, one may wish to group peptides derived from similar sections of a protein having internal repeats (e.g., pool I may contain 30 the first several peptides from the beginning of each ~O :~ lin fold in an antibody).

WO95118148 1802~ PCTII~S93/11703 T cells are obtained from a subject, and are contacted with the peptide pools. For of T cell epitopes specific to a given pathogen, one should select subjects who have been exposed to the pathogen, either as the result of infection or ~7 Subjects who have recovered from infection may provide the best (most protective/diagnostic) 5 epitopes. The T cells may be either fresh or frozen, but are not cultured or cloned prior to assay in order to preserve the natural l" ' ' ' of T cells having different antigen recep-tors. The T cells are preferably obtained from peripheral blood as PBMCs, and are separ-ated from erythrocytes and poly .1' I cells by ~^~ntrifil~ or illl..,r~ r-activated cell sorting. The T cells and peptide pools are placed together in a plurality of 10 wells or culture dishes, and the response ~l~tî.rr^in~d by r~ ~ ~ binding directly (e.g., by r~ of a labeled peptide) or through, ;t~ y (typically measured by 3H-T
uptake after culture). The cells are preferably cultured in autologous serum, in the absence of xenogeneic or pooled serum, to reduce ~ hol. ' responses. T cells which recognrze one or more of the peptides respond by mitosis, clonally expanding the number of specific T
15 cells ~c~ O g, the peptide ~ A~ One may optionally add autologous APCs, either as live cells or as fixed cells, or one may culture the T cells in wells having an autologous MHC protein bound to a support, for improved antigen presentation.
Once T cell epitopes have been fl^t,rrni-^ l, they may be employed as reagents in T
cell-based cc~yS Such assays are advantageous over antibody-based assays because20 they do not require that the subject have already mounted a detectable antibody response to the antigen. The T cell (TH) response to an antigen necessarily precedes the B cell (antibody) response: thus, the T cell response may be detected earlier than the antibody response. Diagnostic assays may be performed with a much smaller sample, because it is not necessary to scan the entire length of the antigen once the epitopes have been ~' ' 25 It is likely that some antigens will exhibit different epitopes in different individuals, based on the l~ uo~ y of the MHC proteins which present the antigen to the immune system.However, one may include a number of T cell epitope peptides in each assay~ For example, one may screen a subject's PBMCs against a pool comprising all known HIVsF2 gpl20 T cell epitopes Response to any peptide in the pool may be counted as a positive response.
.

-Wo 95118148 ~,~ PcrlUSs3/11703 The response may be detected by a variety of methods. The presently preferred method is to culture PBMCs in contact with the peptide pools, followed by pulsmg with 3II-T to determine T cell ,ululif~ iull. One may also detect mitogenic effects by other means, for example, by monitoring the increase in ' k: 2 mRNA using PCR, and the lilce.S Alternatively, one may label the peptides and detect binding directly to the T cell antigen receptor. For example, the peptides may be labeled with fluorescein.
Immune responses l~birds and mammals are believed to require interaction betweenT cells and B cells, where both cell types are activated by contact with an dlU~JIU,U
antigen. Accordingly, effective vaccines should include both B cell and T cell epitopes.
Suitable vaccine î~ ' are k-nûwn in the art for use with the epitopes of the inventiûn.
See for example, EP 399 843, which discloses an adjuvant emulsion C. xamples The examples presented below are provided as a further guide to the lulduliLiu,~ of ordinary sill in the art, and are not to be construed as limiting the invention in any way.
Example I
(M. boYis l:h A) In this assay, the stimulation of T cells is detected by their proliferation after exposure to the test ~UIIIUU ' It will be ~.,u,u" .' that other assay methods can be used without violating the essence of the invention.
Autologous serum was obtained after dt;rlb of 50 to 100 mL freshly-drawn human venous blood by gentle agitation with 3-5 g of sterile acid-washed glass balls (5-8 mm diameter) for at least 10 min. Serum was collected from above PBMC bands after density ~5 interface ~ '' Autologous plasma was collected from above the PBMC band after density interface ~ ilu~ io~l of ~ tj~ o~l ' whole blood. Group AB human plasmas, ~nti~ g ' ' with either acid-citrate-dextrose or citrate-~ dextrose (ACD, CPD), were a gift from (~ w~lLI- Serum T ~hor~nri~-~ (CSL, Ml~!ho~lrn~. Australia). Plasma was converted into serum by the addition of 10 mM CaCI2 and I IUlmL (final) of human thrombin (CSL) to plasma prewarmed to 37C, followed by vi~orous agitation for 5-10 min.

.. . . . . . ... .. . . .. . . _ .. . .. .. . . . . . . . .. . ... . ... .....

WO95118148 18~87 PCT/US93ill703 The mixture was allowed to stand at room temp. for 60 min and the ~ serum was collected by ~ r O '- at an RCF of 20000 at 4C for 20 min. Sera or plasmas were l~Ll Y ' at 56C for 45 min in a water bath.
PBMC were obtained from h. IIA I ;~; ' ' '~ or ~ whole blood from a healthy 5 adult volunteer donor panel Altematively, screened buffy coats from blood ~A~ A I ~
with CPD, kindly supplied by the Red Cross Blood Bank (MP3hournP, Australia) were used Blood was diluted slightly to between 1:1 and 2 5:1 (50% to 72% (v/v) whole blood) with incomplete medium and underlayed with Ficoll/Paque (Pharmacia LKB, Uppsala, Sweden) in a 50 mL ~ùly~u~)yl~ . centrifuge tube in a final ratio of 2:1 diluted blood:Ficoll/Paque A
band of PBMC was isoldted at the interface by ~ ru6d~iu., at an RCF of 450 for 25 min When autologous serum was required, it was recovered from above the PBMC band PBMC
from the band were washed twice by ~ in "il~u-~ Jki~;" medium containing about 10% (v/v) added autologous serum The first wash was at an RCF of 450 for 15 min to ensure efficient pelleting of cells from the medium containing residual Ficoll/Paque, wbile the second was at an RCF of 150 for 10 min to minimize platelet and debris ~ "; -~
Cells were suspended in complete medium containing autologous serum or pooled human serum and viable cells were counted on a ll~,..,o~yLu.~ ,r using trypan blue dye exclusion.
All overlapping 12-mer peptides of the protein MPB 70, produced by the bacteriumMycobacter~um bovis, were ~ ;d. This protein is one that is known to stimulate Tcells, and has the fol~owing primary sequence:
Gly Asp Leu Val Gly Pro Gly Cys Ala Glu Tyr Ala Ala Ala Asn Pro Thr Gly Pro Ala Ser Val Gln Gly Met Ser Gln Asp Pro Val Ala Val Ala Ala Ser Asn Asn Pro Glu Leu Thr Thr Leu Thr Ala Ala Leu Ser Gly Gln Leu Asn Pro Gln Val Asn Leu Val Asp Thr Leu Asn Ser Gly Gln Tyr Thr Val Phe Ala Arg Thr Asn Ala Ala Phe Ser Lys Leu Pro Ala Ser Thr Ile Asp Glu Leu Lys Thr Asn Ser Ser Leu Leu Thr Ser Ile Leu Thr Tyr IIis Val Val Ala Gly Gln Thr Ser Pro Ala Asn Val Val Gly Thr Arg Gln Thr Leu Gln Gly Ala Ser Val Thr Val Thr Gly Gln Gly Asn Ser Leu Lys Val Gly Asn Ala Asp Val Val Cys Gly Gly Val Ser Thr Ala Asn Ala g wo 95118148 2 ~ 8 ~ ~ 8 ' PCT/US93/1 1703 Thr Val Tyr Met Ile Asp Ser Vai Leu Met Pro Pro Aia The 152 peptides were sy..~ ,,~od in 4u~1., r~' ' on plastic rods according to the S methods disclosed im WO90109395. The base-labile moiety Lys-Pro was i.. ~,ull ' at the carboxy terminal of each peptide, and the amimo terminal amine group was acetylated. The peptides were cleaved from the rods into 150 ~L of 0.1 M 1,;~ buffer at pH 8.2, and the solutions of each different peptide were combined into pools.
Fourteen pools of peptides were prepared. Each pool consisted of 11 adjacent 10 overlapping 12-mer peptides. Thus, Pool #I consisted of the 11 peptides IGDLVGPGGCA-EYA to IlYAAANPTGPASV (where the superscript indicates the residue within the MPB 70 sequence above). Pools were made by mixing 200 ,~L volumes of each of the peptide prep-arations. To each peptide pool was added 1.8 mL of RPMI-1640 cell culture medium. Each assay was carried out in 48 replicates. Ly I ' jt..f. were separated from whûle blood (50 15 mL) and ~ ,if~l in RPMI complete medium at a cell, of ~I,lu 'y 1.07 x 106 cells/mL. In the assay, 180 ~lL cell suspension was added to 20 ,uL of peptide pool in ~ ,lu~ ullul_ plates and incubated for 7 days at 37C in a 5% CO2 dllllO~ ;. Six hours before harvest, the cultures were pulsed with I ~Ci of 3H-T. The cells were har-vested, and the amount of 3H-T , ' measured in a liquid 'l~ti~n counter (IKB
model 1205 Betaplate). The positive control was 20 ~L of a 100 ,~bg/mL solution of MPB-70, and the negative control was 20 ,uL of cell culture medium.
A threshold point was calculated based on the probability that no assay result would exceed the threshold by chance (on a two-tailed test). This procedure was repeated until all assay results below the threshold were included in the "background" estimate and change was made in the estimate of the threshold. Thus, when the data were submitted to this procedure, the lowest 615 assay results (of 768 total) yielded a mean of 3170 cpm (standard deviation 1420) and gave a threshold estimate of 7700 cpm. Thus, from this assay, we would not expect any assay result to return a value greater than 7700 cpm if there was nû
specific proliferation of the ry, ' yl~s. In this assay, 153 results returned values greater than this. Pools I and 2 returned 33 and 40 positive results (out of 48 assays), respectively, as shown in Table 1:

~?18o~87 TABLE 1: Positive Results per Pool Pool Number of Mean cells Frequency 5 P~ iYes per 1~cc~y (I in.. ) 33 1 .163 172,000 2 40 1.792 111,600 3 4 0.087 2,298,900 10 4 5 0.110 1,818,200 S 3 0.065 3,076,900 6 0 <0.021 >9,523,800 7 3 0.065 3,076,900 8 3 0.065 3,076,900 15 9 2 0.043 4,651,200 10 1 0.021 9,523,800 I I 0 < 0.021 > 9,523,800 12 2 0.043 4,651,200 13 3 0.065 3,076,900 20 14 6 0.134 1,492,500 Control (+)48 >3.871 <52,700 Control (-)0 <0.021 >9,523,800 The analysis indicates that only pools I and 2 contain peptides which ,, ~
stimulate the growth of T cells ir~ the blood obtained from this particular donor.
B) The assay was repeated with 20 ~L of the individual peptide IJlC~dl~liUIls that were pooled to make pools I and 2. In this case each assay was carried out with 12 rep-licates. The results are shown in Table 2:

Wo 95/18148 ~ Pcrluss3lll7o3 ?~
TABLE 2: Assay of Pools I and 2 Peptjde Number of Mean ceUs Frequency pn~itiVeS perassay (I in.. ) 5 1 GDLVGPGCAEYA 0 <0.087 >2,300,000 2 DLVGPGCAEYAA 0 <0.087 >2,300,000 3 LVGPGCAEYAAA 0 <0.087 >2,300,000 4 VGPGCAEYAAAN I 0.087 2,300~000 5 GPGCAEYAAANP 0 <0.087 >2,300,000 10 6 PGCAEYAAANPT 0 <0.087 >2,300,000 7 GCAEYAAANPTG 1 0.087 2,300,000 8 CAEYAAANPTGP I 0.087 2,300,000 9 AEYAAANPTGPA 4 0.405 493,800 10 EYAAANPTGPAS 5 0.539 371,100 15 11 YAAANPTGPASV 4 0.405 493,800 12 AAANPTGPASVQ 4 0.405 493,800 13 AANPTGPASVQG 3 0.288 694,400 14 ANPTGPASVQGM 6 0.693 288,600 15 NPTGPASVQGMS 4 0.405 493,800 20 16 PTGPASVQGMSQ I 0.087 2,30Q000 . 17 TGPASVQGMSQD 2 0.182 1,098,900 18 GPASVQGMSQDP 0 <0.087 >2,300,000 19 PASVQGMSQDPV 0 <0.087 >2,300,000 20 ASVQGMSQDPVA 0 <0.087 >2,300,000 25 21 SVQGMSQDPVAV 1 0.087 2,300,000 22 VQGMSQDPVAVA 0 <0.087 >2,300,000 Positive Control 12 >2.485 <80,000 - - Negative Control 0 <0.087 >2,300,000 The results suggest that there are two peptides that have a major stimulatory effect on the T cells. These are peptide 14 (ANPIGPASVQGM) and peptide 10 (EYAAANErrGPAS), although other peptides that share parts of the sequences of these peptides do have a partial y effect.
Example 2 (lIerpes Simplex Virus Dr~
Peripheral blood 1,~ .""~ , cells (PBMCs) were collected from ten donors as described below, and fro~en until use.

wo 95118148 218 0 ~ 8 7 ~ , PCT/US93111703 PBMCs are prepared and preserved from 45 mL of whole blood. The cells should not be exposed to bovine or other xerosera. Replicate numbers ( > 24) are necessary to assure reas-onable confidence in the value calculated. The number of cells seeded per well depends on 5 the expected frecluency of responding T cells in the l~l r~
Fresh l . ' whole blood (45 mL) is partitioned in three 50 mL conical centrifugetubes, then diluted with 10 mL of PBS warmed to room t~ y~ lul~. Tbe diluted blood is carefully underlaid with 10 mL Ficoll, taking care not to mix the interface, then cr..,l.;rl~ d in a swinging bucket rotor at 400 x g for 20 min at 20C with the brake off, between 18-22C. The serum is aspirated from above the 1~, ' ~ yl~ band and the cells removed from the interface, taking care to avoid removing material from the Ficoll layer (which contains granulocytes). Each band should be collected in about 5 mL. The PBMC from the three tubes are combined into a 50 mL centrifuge tube and diluted with 4 volumes of PBS, then ~. -.;r,. ;~l at 60~100 x g for 8-10 min at 20C with the brake on. The ~u~are decanted, and the cell pellets .~ . ' ' in 10 mL PBS. The cells are transferred to a 15 mL tube and centrifuge as before. The PBMC are washed one more time with 10 mL
PBS. Cells are ,~ ~l.r ~ l~i in 10 mL RPMI CM-I % PHS and counted in a hemacytometer using trypan blue exclusion to estimate viability. If a significant red cell, r is evident, dilute the PBMC to be counted in Turk's solution. The cells are pelleted again, and n~ .l for frc~ezing in 1.0 mL of RPMI CM containing 20% PHS. RPMI CM (1.0 mL) containing 20% PHS and 20% DMSO is added dropwise while swirling the tube, and the ceUs transferred to two Nunc freezing vials. The vials are placed on ice until all are filled, then frozen at -70C until used.
Assay plates are prepared before thawing the PBMC. Use 96-well U-bottom plates for cultures seeded with ~20,000 cells/well; 96-well V-bottom plates for cultures seedcd with <20,000 cells/well. Antigens are diluted in Iymphocyte basal medium to lOx the final 1;."" and are added in 20 ~L to the wells. The PHA is added 48 hours before the - assay is completed. Basal medium alone is added to the "cells alone" wells. Vials of frozen PBMC are retrieved from the liquid N2 and kept on dry ice until they ready to be thawed.
Six to eight samples can be processed at a time. Conical centrifuge tubes (15 mL) and an , wo 95/18148 . PcT/uss3111703 ?..~Q~
equal number of 60 mm tissue culture dishes labeled with the sample ID are set up.
Lymphocyte wash medium (S mL) is added to each of the tubes and dishes, and vials ~ullc~o~iJ~ , to the samples to be processed placed in a 37C water bath. The viais are removed from the water bath before each sample is completely thawed, the outside rinsed 5 with ethanol, and the contents placed in it's respective dish. One mL of medium from the IS
mL tube is used to rinse the vial. Tlie cell suspension is transferred from the thawing dish to the centrifuge tube, and the dish rinsed with the medium used to rinse the vial. The PBMC
are pelleted in a tabletop centrifuge at 900 x g for 8 min, and the pellets washed twice with 10 mL of wash medium, then Ir.~ --lf,l in S mL of Iymphocyte culture medium. The10 cells are counted after the second wash and the volume of medium and cells needed to set up the plate calculated. To seed 96 wells with 180 ~L of cell suspension/well, 18.5 mL of cell suspension is needed. The number of cells seeded per well depends on the expected responder frequency. If the expected frequency is low (e.g. I responder in 500,000 PBMC), 105 cells/well is a good starting point. For expected r . ' of I in 20,000 to 1 in 10,000 about 20,000 cells/well should be used. The object is to achieve roughly half of the antigen-containing wells scored as positive. The washed cells are diluted such that the number of cells/well will be delivered in 180 ~L to the U-bottom plates. When V-bottom plates are used for high-frequency assays, the cells are ~r~ at 1-3 x 105 cells/mL
and tlle appropriate volume added. As little as 20 ,bL of cells can be seeded in V-bottom wells. However, in this, ~ ~ , the cells should be diluted in medium containing anti-gen rather than adding the antigen to the wells prior to thawing the PBMC.
The plates are incubated in a humidified 7% CO2 incubator. Two days before the pulse, PHA is added to the d,UIU~UiJl;ai~ wells in 20 ,~L of basal medium. Early on day 6, 20 ~bL of 3H-thymidine diluted to 25 mCi/mL in basal medium is added to each well using a ' ' pipettor, and the plates returned to the incubator. After six hours, the plates are harvested with a Cambridge 2800 harvester using program 0. Alternatively, the plates can be frozen and harvested at a later date. The filter mats are dried for several hours or overnight before sealing them into their bags with 10 mL of 5(~intill~imn fluid. The filters are counted in the ,~-plate counter using the a,u,U.ulu protocol.

Wo 95118148 2 1 ~ ~ 2 ~ 7 PcrluS93/11703 .
Anti~.n Pn ~ tion Peptides 13mers were prepared in 4ua~ . ' on plastic rods according to the methods disclosed in W099109395, as described in the ~xample above, based on the amino acid sequence of the Herpes simplex virus 2 antigen gD2:
5 Lys Tyr Ala Leu Ala Asp Pro Ser Leu Lys Met Ala Asp Pro Asn Arg Phe Arg Gly Lys Asn Leu Pro Val Leu Asp Gln Leu Thr Asp Pro Pro Gly Val Lys Arg Val Tyr His Ile Gln Pro Ser Leu Glu Asp Pro Phe Gln Pro Pro Ser Ile Pro Ile Tbr Val Tyr Tyr Ala Val Leu Glu Arg Ala Cys Arg Ser Val Leu Leu His Ala Pro Ser Glu Ala Pro Gln lle Val Arg Gly Ala Ser Asp Glu Ala Arg Lys His Thr Tyr Asn Leu Thr Ile Ala Trp Tyr Arg Met Gly Asp Asn Cys Ala Ile Pro Ile Thr Val Met Glu Tyr Thr Glu Cys Pro Tyr Asn Lys Ser Leu Gly Val Cys Pro Ile Arg Thr Gln Pro Arg Trp Ser Tyr Tyr Asp Ser Phe Ser Ala Val Ser Glu Asp Asn Leu Gly Phe Leu Met His Ala Pro Ala Phe Glu Thr Ala Gly Thr Tyr Leu Arg Leu Val Lys ne Asn Asp Trp Thr Glu Ile Thr Gln Phe Ile Leu Glu His Arg Ala Arg Ala Ser Cys Lys Tyr Ala Leu Pro Leu Arg ne Pro Pro Ala Ala Cys Leu Thr Ser Lys Ala Tyr Gln Gln Gly Val Thr Val Asp Ser Ile Gly Met Leu Pro Arg Phe Ile Pro Glu Asn Gln Arg Thr Val Ala Leu Tyr Ser Leu Lys ne Ala Gly Trp His Gly Pro Lys Pro Pro Tyr Thr Ser Thr Leu Leu Pro Pro Glu Leu Ser Asp Thr Thr Asn Ala Thr Gln Pro Glu Leu Val Pro Glu Asp Pro Glu Asp Ser Ala Leu Leu Glu Asp Pro Ala Gly Tbr Val Ser Ser Gln ne Pro Pro Asn Trp His ne Pro Ser Ile Gln Asp Leu Ala Pro His His Ala Pro Ala Ala Pro Ser Asn Pro Gly Leu Ile Ile Gly Ala Leu Ala Gly Ser Thr Leu Ala Ala Leu Val Ile Gly Gly Ile Ala Phe Trp Val Arg Arg Arg Ala Gln Met Ala Pro Lys Arg Leu Arg Leu Pro His Ile Arg Asp Asp Asp Ala Pro Pro Ser His Gln Pro Leu Phe Tyr wo9sll8l48 ~ Pcr/US93/11703 --Twelve pools of peptides were prepared. Pool I contained 12 peptides, pools 2-11contamed 10 peptides, and pool 12 contained 8 peptides. Pools were made by mixing 200 ,~L volumes of each of the peptide ,U~ ,dldliulls. To each peptide pool was added 1.8 mL of RPMI-1640 cell culture medium. Each assay was carried out in 48 replicates. In the assay, 180 ~4L cell suspension was added to 20 ~L of peptide pool in microculture plates.
Results:
Each pool was scored for strength of response (none, weak, moderate, high) for each donor. Pools 1-3 and 7-12 exhibited weak to 2ero response. Pools 4-6 exhibited at least moderate responses for 7, 6, and 5 of the donors, respectively.
For three of the donors, samples were retested against individual peptides from the responding pools to determime which peptides were ~ ,Uu...il)l~ for the observed activity.
Donor A responded to peptides in the following regions: l9GKNLPVLDQL, 97IAWYRMGDNCAIPITV, l6lAGTYLRLV, and 202LTSKAYQQG. Donor C responded to peptides in the following regions: l36u~ Kw~YDSF, and l701NDWTEITQFILE.
15 Donor D responded to peptides in the following regions: l33~KW~KW~YDSFS-AVSEDNLGFLMHApAFETAGTYLRLVKINDW~ ~HRARASCKYAL (not resol-ved), and 232ALYSLKIAGW~GPKP. Samples from Donor D were additionally tested for CD8+ T cell epitopes. The following CD8 epitope was flf~tf ' 307APAAPSNPG.
T ~ ,ly, the CD8 epitope occurs im a region which is different in HSV-I and HSV-2, 20 and may account for the fact that immune responses to HSV are type-specific.
xample 3 (Human r ~ ~- -y Virus re~ ) Eight subjects were selected from ~Jdl~ Jdlll~ in a clinical study to examine a recom-25 binant HIV vaccine based on HIVsF2 gpl20. The subjects were vaccinated with HlVsr2gpl20 in an adjuvant ru~ ldLio.. on day 0, and boosted at one month and six months. Some subjects received placebos. Samples were obtained and treated as described in Example 2 above, both prior to vaccmation and two weeks following the last boost.
Peptides were prepared as described in Example 2 above. The peptides were 15rners, offset by two residues, spanning the length of HIVsF2 gpl20 (excluding the 30 residue signal , . _ .. ... .. . . .. .. _ _ WO95/18148 2l8D287 PCTIUS93/11703 sequeDce). The gplZ0 sequence is shown in Figure 1. The peptides were grouped in pools of 13 peptides, for a total of 18 pools. The results are shown in Table 3.
TABLE 3: Assay of H/V Pools Peptide Negative Ab Positive Ab Pool ~ ~ pre yost pre post r3 343-381 0/2 0/2 1/6 3/6 gpl20 31-509 0/2 0/2 0/6 616 30 In Table 3, "Negative Ab" indicates the two subjects who did not produce antibodies to gpl20 detectable by ELISA (and thus were probably placebo recipients), while "Positive Ab"
indicates the six subjects who did produce gpl20 antibodies. Full length (31-509) gpl20 was included as a positive control.
- The results indicated that pools 3 and I I probably contain the T cell .~ for 35 ~VSF2 gpl20, based on the increase in response between pre-vaccination and post vaccrnation tests. Pools 3 (peptides 27-39) and 11 (peptides 121-133) were analyzed for five subjects to deterrnine which peptides were ~ )u..~ for the observed activity. Pools for Woss/lsl4s ~ Pcrluss which responses were obtained both before and after vaccination probably indicate cross-reactive Witll other pathogens. The results are shown in Table 4.
TABLE 4: Assay of Pools 3 and 11 Peptide Seguence Response post-vaccination 27 83vvLGNvTENFNMwKN 0/5 28 8sLG~VT~N~ 1/5 29 87NVTENFNMWI~NNMVE 2/5 891E~FNMW~VEQM 5/5 31 9 INF~1MWKN~EQMQE 5/5 39 l07IISLWDQSLKPCVKL 2/5 121 29lLNESVAINCTRPNNN 1/5 124 297INCTRPNNN~SIY 2/5 125 299CTRPNNN~RKSIYIG 0/5 126 30lRPNNNTRKSIYIGPG 1/5 127 3o3NNNTRKsIyIGpGRA 1/s 131 3llYIGPGRA~ 4/5 132 3l3GPGRAFl~TTGRIIGD 4/5 133 315GRA~H ~ wLGDIR 5/5 The results (I,.. ,~ that peptides 30-33 and 129-133 include T cell (I` lf 'llli'~ for EIIVsF2 gpl20 wo ss/l8l4s pcTluss3lll7o3 Examplç 4 A. General Batches of RPMI-1640 medium were obtained from the t~ . ~lth Serum T~h~lra~rjP~ ("CSL", Parkville, Victoria, Australia), Gibco T ~nrltr,riP~ (Grand Island, S NY), Flow T. I (North Ryde, Australia), or Hazelton Inc. (Lenexa, KS). Eagle's minimal essential medium (MEM), Dulbçcco's modifled ME~ (DMEM), Monomed serum-free liquid medium, penicillin/s~ ,A..y~ kanamycin sulphate, nçomycin 'i ' /I,olymyxin B sulphate, L-glutamine, sodium bi~.dlll~ ' and HEPES
buffer were supplied by CSL. Ciprofloxacin was kindly donated by Bayer Australia10 (Sydney, Australia). ~rnrl~nnrlp~rl~ culture medium I . nn was RPMI-1640 r~ with L-glutamine (2 mM), sodium bil,dlL ' (2 g/L), gentamicin (50 ~g/mL) and HEPES (5 mM). "Complete" medium was incomplete medium ., ' 1 with 10%
human serum.
Autologous serum was obtained after ~lrril~ ., of 50 to 100 mL freshly-drawn human 15 venous blood by gentle agitation with 3-5 g of sterile acid-washed glass balls (5-8 mm diameter) for at Içast 10 min. Serum was collected from above PBMC bands after density interface ~l6aLiul~. Autologous plasma was collçcted from above the PBMC band after density interface ~".Llirb6dio-- of ~nfirr~glll ' ' whole blood. Group AB human plasmas, ~ntirr~ with either acid-citrate-dextrose or citrate-phosphate-dextrose (ACD, CPD), 20 were a gift from CSL. Plasma was converted into serum by the addition of 10 mM calcium chloride and I IU/mL (final) of human thrombin (CSL) to plasma pl~.wa~ ~l to 37C, followed by vigorous agitation for 5-lO min. The mixture was allowed to stand at room L~ . ci for 60 min and the ~,. I IA..I serum was collected by ~. ..";r,.,,~""ll at an RC~
of 20000 at 4C for 20 min. Sera or plasmas were heat-inactivated at 56C for 45 min in a 25 water bath.
PBMC werç obtained from ll~dl i,li~,d or dciribl ' whole blood from a healthy âdult volunteer donor panel. Alternatively, scrççned buffy coats from blood ~ntirrl~ll ' with CPD, kindly supplied by the Red Cross Blood Bank, Melbourne, Victoria, were used. Blood was diluted slightly to betweçn l:l and 2.5:1 (50% to ~2% (v/v) whole blood) with 30 incomplete medium and underlayed with Ficoll/Paque (Pharmacia LKB, Uppsala, Sweden) in . , ,, . ,,, _ _ , ,, ,, _ _ _ , , . , _ , _, WO 95118148 ?~ o~ PCT/US93/11703 a 50 mL poly,ulu,uyl~ , centrifuge tube in a final ratio of 2:1 diluted blood:Ficoll/Paque.
band of PBMC was isolated at the interface by ....1. ir,.~,.,il . at an RCF of 450 for 25 min.
When autologous serum was required, it was recûvered from above the PBMC band. PBMC
from the band were washed twice by I ~.,I.;r.~".linn in "incomplete" medium containing about 10% (v/v) added autologous serum. The first wash was at an RCF of 450 for 15 min to ensure efficient pelleting of cells from the medium containimg residual Ficoll/Paque, while the second was at an RCF of 150 for 10 min to minimize platelet and debris, IIII~
Cells were suspended in complete medium containing autologous serum or pooled human serum and viable cells were counted on a ~,l..ouylu...~ ,. using trypan blue dye exclusion.
Tetanus Toxoid (Tl~, and Purified Protein Derivative (PPD) from BCG cultures were supplied by CSL, MPlh~llrllP, Australia, and were used at 0.1-10 Lf/mL and 0.1-10 ~g/mL
respectively. TT was dialysed to remove thiomersal preservative before use.
B~ u~u,ul ' ~ vated zonally purified A/Shanghai/l 1/87 (H3N2) influenza virus suspension was also supplied by CSL. Concanavalin A (Sigma, St. Louis, MO) was used at 2-10 ~g/mL. T~f~ Herpes Sirnplex virus type 2 glycoprotein B was kindly suppliedby Chiron Corp. (Erneryville, CA).
Single ul~ol~-scale resin peptides were made on a Milligen 9050 synthesizer (Milligen/Biosearch, Rllrli-~"f~m, MA) and purifled by reverse-phase HPLC. Multiple peptides were made on pOl~ pins using the Multiple Peptide Synthesis System (Chiron l\' '(r ~ Clayton, Australia). Solution phase peptides were generated bycleavage from the pin at neutral pH of a Iysime-proline ester linker to forrn a dik~,tvfuiu~ e (cyclic dipeptide) moiety at the carboxy terminus of each peptide.
In the "standard" 1~ u~uL~ iu~l assay, PBMC (2x105 per well) along with antigen or mitogen were added to 96 well U-bottom tissue culture trays (Nunc, Roskilde, 25 Denmark) in a funal volume of 200 ,uL complete medium per well. Cultures were incubated at 37C in humidifled 5 % CO2 in air for various lengths of time. The stated amount (usually 0.25-1.0 ~Ci) of tritiated thymidine (3H-TdR) (Arnersham T..'l .l -' jl."A;, Bucks, UK, or ICN Rj~mPAi~`'AI~, USA, specific activity 40-60 Ci/mmole) was added and the plates 1~ ' ' for the stated time before harvesting the DNA onto glass fi~re filterrnats and counting in an LKB 1205 Betaplate liquid 5, infill~fion counter (LKB, Turku, Finland).

... .. . .. .. . . ... _ _ .. _ .. . ... .. .. ... _ . _ _ . ., . . . _ . .

Wo 95/18148 1 8~2$ 7 PCT/uss3lll7o3 V-bottom 96 well microtiter culture trays were from Greiner GmbH, Nurtingen, Germany.
Flat-bottom 96 well microtiter culture trays were from Nunc, Roskilde, Denmark.
Tests were performed in as many replicates as the cell yield and size of the blood donation from each donor allowed. A minimum of 12 replicates per ~iA,U~ group was 5 generally used. Results are expressed as: the mean i SD of the cpm of the replicates; as stimulation indices (mean cpm of stimulated cultures/mean cpm of Cells Alone controls); or ~~ the replicates were scored as positive (responding) or negative (I~uru-,~r ~ ~) based on a cutoff value for cpm of the iu~ulpl ' 3H-TdR. For the calculation of the cutoff, data from all e,~,u, I groups was pooled for analysis.
10 B. ~
Seven lots of RPMI-1640 medium, and two other base media (MEM and DMEM), were used to make up "complete medium" without antibiotics, ~u~u~ with identical additives including 10% (v/v) of a single batch of pooled human serum. They were then tested in ululirulaLi.ll assays with PBMC from one donor, in response to antigenic 15 stimulation with PPD (Table 1). The ch,qrq~t~.ticti--~ sought were congruous, low ba~,h~;luu levels with strong antigen-specific ~lulir~,ld~ . RPMI-1640 media tested included five different brands (three liquid and two powdered) and different batches of two brands. We also tested MEM and DMEM. A serum-free medium, "Monomed~', was tested without sup-I ' with serum. We found that some media gave high mean bd~h~luuild (cells 20 alone) p~ulir~ldliull, and some gave high frequency occurrence of ~ proliferatio in the ba~h~;luulld group (media G and F, Table 1). Some of the ba~,h~lL ' group wells which underwent -r ' ~ proliferation gave counts as high as the specific responses (data not shown). Brand A liquid medium was found to be superior to all others, and further testing of batches of brand A confirmed its consistent ~ilrul (results not shown).
25 MEM and DMEM both supported adequate levels of specific lululir ld~iull, but were not the most suitable due to high or uneven ba~h~;luull.lq.
In CSL Monomed serum-free medium, antigen-specific responses were markedly reduced compared to RPMI-164û containing 10% pooled human serum.
Negative control cultures of PBMC from some donors were found to give ~nq~crrtqhly 30 high rl~u, ~,.lUiCs of " . " proliferation in medium containing the antibiotic wo 95118148 PCTiU593/11703 penicillin/~LI~Iu~u~lly. ;ll but not in medium containing O With some donors, responses to amtigen were also heightened in the presence of these antibiotics.
Several other antibiotics were then tested for their suitability: ~ hanamycin, neomycin/polymyxin, and ~ipll rl~ Gentamicin was chosen due to its low toxicity and S absence of ' y effect on controls, as well as its stability during culture.
As shown above, serum is a vital component of media for ,u~ulir~ hal~ assays on PBMC.
Xuman AB sera are widely used for the purpose. However, the selection of reliable sera also requires a screening process, as illustrated by the following data. ACD . ~human plasmas from blood group AB donors were converted to serum and each separate 10 donation tested as a medium , I ' in antigen- and mitogen-driven ululi~ iull of three lots of human PBMC. Serum was graded as suitable for use on two criteria. Firstly, eacl lot of serum should provide a low, even background with all three test PBMC. Secondly, the serum should provide good growth (>3x mean of ba~hOIuulld wells) of both themitogen and antigen stimulated cultures of at least two of the three lots of PBMC. From this and the results with the other two PBMC, sera 40, 44, 46 and 50 were excluded. Sera 44 and 50 were excluded because they did not support proliferation in the other two PBMC
tested. Serum 40 did not support antigen-driven ~uliru~dLiull in any of the PBMC even though il supported mitogen-driven proliferation. Thus, screening human AB sera can be a large u..l~.L~hi..g requiring careful evaluation. Commercial human AB sera also need to be 20 tested for this purpose as batches vary enormously in suitability.
Sera which fail the selection criteria ("poor" sera) would still be usable within a serum pool if they simply lack factors present in "good" sera, especially if "good" sera are u~ u~!plicid with such ~ However, if growth ~u,utJI~;llo factors are present in poor serum, no benefit will be gained by mixing with good serum. Three good sera and 3 25 poor sera were tested separately, and as pools, in antigen-driven ,u~ulir~ iul. assays with four donors. One of the poor sera, P3, was designated "poor" because it generated frequent ~ullkuleuu~ proliferation in controi cultures rather than because of lo~ support for ,UlUlif~ Liull. When all three good sera were pooled (Gl.3), results were similar to those when each was tested separately. The ,ululir~ Liul~ ~ p,uulLillg capacity of serum P3 was 30 reduced when mixed with the other poor sera (Pl.3). Lihewise, when the poor sera were WO95/18148 21 8o28 7 PCT/US93/11703 pooled with the good sera the antigen-driven proliferative response was greatly reduced by with the pool of good sera. These results mdicate the presence of ~UIJ~ D~ C
factors in certain sera rather than a growth factor deficiency.
Autologous serum from ~' ~' ' blood was found to be at least as good a medium 5 ~ as screened pooled serum. -In most cases media contaiming autologous serum gave higher stimulation indices than media containing pooled serum. This was particularly so for antigen-driven ~lulir~,ldliul~ rather than mitogen-driven proliferation. The practical advantage of usimg autologûus serum is that it requires no IJI~,D~I~ Ig, is very easily obtained in sufficient quantity for the r~ : at hand, and requires no additional10 processing to remove ,",lil IJ~ or fibrinogen. Autologous serum from a previous bleed of the same donor can also be used.
Variation in autologous serum c, had no strong effect on antigen-driven Ii~ldliUll over the rdnge of S to 20% serum. All four c,, . .,~ i.. tested supported both whole antigen- and peptide-driven proliferation. There was no consistent trend in the mean cpm of wells scored as negative (b~.~k~;.u A.;l). Likewise, the frequency of positive wells and the mean cpm of positives did not change s;~.~ir.~llily as serum ,- dLiun increased.
The effect of time of incubation prior to pulse labeling on ability to detect specific proliferation was determined using complete medium containing 10% autologous serum, 2 x 105 cells per well and U-bottom wells. An incubation time of 3-5 days gave the clearest dirl;.~ idli().. between controls and antigen-stimulated cultures, both for whole antigen and for a peptide. The variation in cpm for control (cells alone) wells increases with time, Witl the most variation occurring after 5 days incubation.
To examine the variation in negative controls ("Cells Alone" controls) over a range of 25 incubation times, four donors' control PBMC were pulse-labeled after one to eight days incubation, using large numbers of replicates. For the first four days' incubation, the mean thymidine incorporation rel~lained fairly constant, but for the next four days there was a steady rise. The coefficient of variation of the mean cpm also increased remarkably from day 5 onward, due mainly to an increasing frequency of wells undergoing A,r 30 I~ulirt;~ iu~. Based on the data thus obtained, and parallel tests with additional WO 95/18148 i PCTNS93/11703 r~
antigen-driven cultures, four days was chosen as the best incubation time for clear ~I-rr between control and stimulated cultures in rl..~ - mapping studies using peptides. It is possible that PBMC with very high rli, . of antigen-specif c precursors may benefit by use of a shorter incubation time (3 days) and conversely that PBMC with a 5 low precursor frequency may benefit by a longer mcubation time (5 days).
To determine if well shape had an effect on PBMC ~uliru,~iul. assays, cells wereincubated in either U- bonom or flat bottom wells for a range of times with various antigens.
The frequeDcy of positives in the No Antigen (cells alone) control, and the frequency of responses to TT, were largely unaffected by well shape. However, cells responded to a T
10 cell ~ peptide from tetanus toxin (P399) more frequently in the U-bottom wells than in flat bottom wells, at both peptide , regardless of the time of incubation.
This may be due to the greater cell-cell contact afforded in U-bottom compared to flat boUom wells. As expected, similar results were found for V-bottom wells, although the latter were not tested as thoroughly as U- and flat bottom wells.
The effect of initial cell number per well on magnitude of the proliferation was measured at a single time point (7 days) after initiation of the cultures. Medium volume was a constan 200 ~L per well. At 1.2 x 104 PBMC per well, no antigen-specific response was .' ' ', possibly due to ~ numbers of antigen-specific precursor T helper cells.
With increasing cell number up to 4 x 105 cells/well, there was a progressive increase in both the mean antigen-driven response and the b~,k~ ' EIigher cell f~ - Ied ~o a reduction in antigen-driven ~Ulir~ probably due to depletion of medium ~
during the seven day incubation period. Therefore, use of cell numbers above about 4 x lOs cells per well may . Ullli~e the detection of specific ~lulirud~iu~. when long incubation times are used.
The effect of initial cell number per well using a variety of well shapes was also determined at a shorter incubation time. PBMC ..,..~...I,.,I;...~C Io6, 3 X 105, and 105 cells/mL, added in volumes of 100, 50 and 20 ~L of medium were used to initiate cultures with between S x 103 cellslwell and 105 cells/well. An incubation period of 4 days was used. The response was expressed as percentage of test wells scored as positive for 30 I~lulir~ liull. Use of V-bottom wells gave the highest sensitivity of detection of WO95/1814~ 218~287 Pcrluss3lll7o3 antigen-specific responses, allowing tesponses to be detected at 5 x 103 PBMC per well for this donor. Under these conditions, a L~ of the results shows that specific pro-liferation in either U-bottom or flat-bottom wells was erratic until cell numbers reached 3 x 104 or 105 cells/well respectively. That this was not simply a lack of responsive T cells is 5 shown by the result with the V-bottom wells.
The effects of varyimg the pulse duration, specific activity, amount, and of the 3H-TdR label on readout of PBMC u-vlir~"~Liull assays were examined.
Mitogen-stimulated PBMC were initiated at 5000 cells/well and pulse-labeled on day 3. The cells i..cu.~,, ' more label into DNA as the ~ . Pll--l~- ron~Pnt~tif)n of 3H-TdR was 10 increased. The amount of label il~,UII~U ' ~ did not reach saturation within the range of ''nnc of thymidine tested, regardless of the specific activity of the thymidine. In both PBMC and a continuous B-cell line, the cpm of ;~ .u~ 1 thymidine increased with increasing specific activity, except for the continuous B-cell line at the highest specific activ-ity tested (48 Ci/mmol). Increases were not directly plUI)Uli' I to specific activity. The 15 cells were able to ~.~,UIIJUI_'e more total thymidine of low specific activity, indicating they were not flooded with thymidine over the dose range tested.
Time course studies of 31I-TdR uptake in ~vlir~ ~ cultures showed a plateauing of illCull~ulll~iull of thymidine after 6h for both PBMC and a continuous B-cell line. This effect was ~LI ' of the specific activity over the range tested, and hence the total thymidine 20 l- - nn Further studies indicated that even after incubation overnight (> 16 h), no further significant increase in label i~,u~,uuld~iOa occurred.
The incubation time at which maximal proliferation occurs is a factor which should be evaluated in any ~ul;rt~ ioll assay. Use of a long incubation time prior to pulse labeling of stimulated PBMC could give the impression that low doses of antigen are more effective than 25 high doses, due to the "overgrowth" of cultures under optimal ctiml~l- conditions.
Conversely, use of a short incubation time may not allow the ~ Jlili~Liull mechanism to proceed to the point where the effects of single antigen-specific precursors could be detected.
An attempt to overcome this timing difficulty was undertaken by trying to maintain available levels of 3E~-TdR in the culture medium. Cells might then :~c-~m~l f^ label progressively as Wo 95/18148 PCT/uss3lll7n3 they proliferate up to the time of harvest. In this manner, the total proliferation of the cultures over the entire incubation time could be measured.
A large number of replicates of antigen-stimulated PBMC were initiated. Each day, a proportion of the cultures were pulse labeled 3 h and harvested. Others were pulsed daily S for 2, 3, 4 or 5 days, beginning after 3, 4, 5 or 6 days of label-free irlcubation. Although u.~,, ' thymidine cpm values were noticeably higher for multiple-pulsed cultures than for single-pulsed cultures, the higher background (Cells Alone) cpm meant that there was little difference in stimulation indices and thus no advantage in multiple pulsing.
The increase in cpm between 3 h and 27 h for cultures harvested after a single pulse, regardless of the day on which the first pulse was given, is consistent with the finding that illCUlllJU14~iUI~ continues up to 6 h after addition of the thymidine.
It is often convenient to store l;.y ' pulsed cultures before harvesting. Replicate cultures were either harvested ' 'y after the pulse label period, stored up to 24 hrs at 4C or frozen for various periods. We found no significant differences in the mean cpm or IS standard deviations. Therefore, if time restrictions or assay size hamper the immediate harvesting of pulsed cultures they can be fro~en for storage of indefinite duration, or held at 4C for I day, provided all plates of a single assay are handled identically.
C. Results ~
The ~ I conditions should not only avoid nonspecific l~ion but aiso provide an environment in which specifically stimulated cells can proliferate optimally. We have developed methods which do not simply look for the greatest magnitude of thymidine u~tJUI~liull or stimulation index in PBMC, but rather for the highest sensitivity and reliability in detecting and counting antigen-specific T helper cells.
Previous workers have noted the importance of culture conditions in obtaining reliable, c~J-udu- ;blc results. Other studies showed that factors such as cell ~ incubation period, pulse label parameters and ~ ;",l of stimulant had interrelated effects on t~
measured tJlUIi~ld~iV.I rate of PBMC. Those studies ~ . .", ~ 1 mainly on the effects of mitogen or allogeneic l ~ion E.J. Hensen et al., Hum T I (1984) 10:95 noted the importance of using a large number of replicates to accurately measure the response to a single antigen. Tl ,u-i 'y, Hensen eraL ~ the effects of random ~ lircl4~iu Wo 95/18148 PCT/USs3/11703 ~18~7 occurring in control cultures, and the inadequacy of the stimulation index as a measure of specrfic ¦JlulirUIdLiUll.
Our results showed that medium-related factors which affected the frequency of IIU~ P.~-specific positives in the Cells Alone control were the base medium, antibiotic, S and serum. Batches of RPMI-1640 medium from different r ' L~l ~ varied remarYably in quality, not only in ability to support growth, but also in their tendency to give rise to antigen-llu~ rlc ~IUIif.,,dLiull. P.T.A. Srhpllp~nc e~ al., Clin Fyn T ' (1968) ~:571 noted, with PHA-stimulated cultures, that the results were not dependent on the choice of base medium, in agreement with our contention that culture conditions in antigen-driven pro-10 liferation are more critical than in mitogen-driven UlUIif~,ld~iull. RPMI-1640 is used because it was specifically designed for human Iymphocyte culture. High b~l~hg~uu~ created by media c, , may be due in part to the presence of various levels of B-cell-stimulatory en~otrn~ , or ~ ~ traces of T-cell- 1 y substances such as bacterial 15 Although most cell culture texts advise the use of penicillin and streptomycin (P/S), we have occasionally found this antibiotic ~ ' unsuitable. Negative control cultures of PBMC from some donors gave ~ 11y high rl~u~".c,h,~ of "`l)" ~ " ,ululif~,aliu in medium containing P/S. With these do~lors, the response to antigen was also heivhtPn~fl The ~,u~ p.ulirc.d~h~.l may be due to penicillin, to which many people are allergic.
20 however this has not been proven. (~Pn~lrnirin was found to be free of such problems and was used routinely.
Serum is the most important single factor in the success of proliferation assays. The rlPtrimPn~l effects of using stored, unscreened serum as well as the differing effects of different serum sources on either antigen or mitogen stimulation have been noted by others 25 bovine serum and other l~ lul~,u.~ sera enhance l,acl~l~. ~ ululif~ld~iull in human PBMC
assays to levels which can obscure antigen-specific responses, whereas selected human sera do not. In contrast, foetal bovine serum has also been found to be inhibitory to human PBMC ulUlir~,ldliu.~ in whole blood culture systems. Despite reports of serum-free media for culture of human PBMC, there has been no general acceptance that serum-free media will wo 95/18148 f8~ PCT/US93/11703 support efficient proliferation of human PBMC, and we also found that serum ' was needed for the serum-free medium rOI we tried.
Selection of suitable human sera is best performed usimg PBMC of several donors whose responses to certain antigens is already well l.l,,~,,., ~. .;,. ~l Sera supporting strong specific s ~lullr~.dLiul~ and low 1,~1~.. ' can then be pooled, aliquotted and stored at -20C or colder. An attractive alternative to screened human serum is autologous serum. PBMC and serum can be recovered from the same whole blood sample in high yield by l:l ;ri~l and dilution of blood with RPMI-1640 medium, followed by density interface, ~ .ir~ as described. This process provides more autologous serum than required for a 10% (v/v) final 10 , r~n in culture medium at the cell densities used. We found autologous serum to be at least as good as screened pooled human serum. Autologous serum can be heat-inactivated at 56C for 45 min without loss of growth I)IJUI~ , qualities.
Our data show that it is unwise to mix poor sera with good sera in an effort to mahe the supply of serum last longer. The inhibitory ~ -r in serum are yet to be defined. It 15 has been reported that high ~r,... ~ l of certain, , ' f , ' . in some sera inhibit PBMC plulif~,ldliull. Best use of good sera can be obtained by using them at j%
(vlv) .
The incubation time and the number of PBMC added per well are l~k.~ factors which must be adjusted toget~er. If input cell, -~n is too high, proliferation at long 20 incubation times can be dramatically reduced, possibly by exhaustion of the nutrient or buffering capacity of the medium, or by inhibition of cells due to high cell densities. Use of too few cells per well requires very high numbers of wells to allow detection ofantigen-specific precursors at biologically significant rl~, (e.g. > 1 per million PBMC). Short incubation times (4 days) give sensitive detection of positives witll low and 25 consistent l; ~.~h~
Maximum sensitivity of detection of ,ulUIir~,.dliul. is vital to calculation of accurate antigen-specific precursor frequency when limiting dilution data is sought. The optimal time for incubation of antigen-driven PBMC proliferation assays using the methods described is 4 days, whereas 2 to 3 days is ~ptimal for mitogen-driven cultures.

wo 95/18148 PCTIUSs3/11703 Cell numbers per well and well shape are also , ' ' factors in antigen-driven PB~C ~ JI;f~,làLiol~. The use of round-bottom wells to improve cell-cell contact allowed the use of reduced cell numbers in ~JlUIil;ldliUIl ~,.,u, Use of V-bottom wells, although useful for improving sensitivity, should be avoided if high cell numbers ( > 100,000 5 cells/well) are to be used, as responses can be inhibited under these conditions.
Exogenous thymidine is quickly assimilated into the ;~ IAI pool and used for DNAsynthesis during the S phase of the cell cycle. Trace-labeling can be used to measure rates of DNA synthesis, provided the thymidine is of low specific activity, <2 Ci/mmol.
However, use of low specific activity thymidine can lead to addition of excess total 10 thymidine, changing the conditions to flood labeling. Use of high specific activity thymidine (40-80 Ci/mmol) results in cytotoxicity, probably due to r~ olr,gir:ll damage. We chose to use trace labeling with a low dose of high specific activity thymidine since the cultures are terminated after a short labeling time (6 h) and the intent is primarily to detect significant differences between IJlUIil;_ld~;.,g and nonproliferating cultures. As thymidine is sold by 15 radioactive corltent, a small dose of high specific activity is the most ~ I way to Iabel. Under these conditions, the dose of thymidime is limiting and il~,UlUI ' ' cpm are thus ulu~Jul~iullal to the rate of DNA synthesis at the time of addition of the label.
As may be expected from the shape of the curve of il~.UlpU ' ' thymidine against time, the use of multiple small doses of thymidine over several days did not significantly enhance 20 the total ...~UI~JUldliUI~ of thymidine by l,U...~)dli~UII with a single dose incubated for 27 h, and any advantage in total ill~UIlUUldliu~ by antigen-stimulated cultures was lost due to increased i",.,.l..,,~i..,~ in the, ' ' controls.
We often find it convenient to store assay plates after the pulse label period. Plates may be placed at 4~C overnight or frozen at -20C for several days without apparent loss or breakdown of the DNA. Plates may be then processed for harvest of DNA and counting of thymidine as time permits.
Although it is wise to test parameters for every PBMC/antigen ~ul ' , the following IC~ apply in most cases.
1. Use screened HEPES-buffered RP~I-1640 with 10% autologous serum or screened, pooled human serum, adding gentamicin if an antibiotic is required, wo 9S/18148 Pcr/uss3/11703 2~ grt 2. Culture up to 2 x 105 PBMC/well in multiple wells (8 or more) in U-bottom microtiter plates or up to 105 PBMC/well in V-bottom plates for 4 days, 3. Trace label with 0.25uCi/well of high specific activity (40-80Ci/mmol) 3H-TdR for the final 6h of the incubation period prior to harvesting the DNA onto glass S fibre filter mats and ll ~ri-)n counting.
4. Analyse the results using an algorithm which takes account of the possible presence of responding wells in the Cells Alone l~olu~,~u,ld;ll~ wells (wells lacking antigen-specific precursors) among the wells to which the test antigen has been added.
Example 5 A. ~Ç~I~
Overlapping ~ ;.lr, for ~ scanning were synthesr~ed with termini consisting of a carboxy-terminal beta-amino-alanine-di~Gto~,;l,~,.~i..~, (~-dkp) group and an acetylated amino terminus. The multipin peptide synthesis strategy was used. Work with several clonal T cell systems has shown that N- and C-terminal-blocked peptides are as efficient or more efficient in T helper cell activation than unblocked peptides, in contrast to cytotoxic T cells. Peptides were cleaved into sterile 0.1 M sodium bil,~.l in 96-well microtitre trays. The purity of .G~,.G~GI.LdLive peptides was assessed using HPLC and was found to be generally >80~. Wells were found to contain an average of 10 mnol cleaved peptide by amino acid analysis.
Bulk peptides:
P399 (Ac-QElYMQTlTYPIS-b-dkp, tt 257-268), P442 (H-EQDPSGATTKSAM-LTNI~(i~ VIl~KNEV-OH, tt 141-171), P443 (H-SVDDALINSTKIYSY-FPSVISKVNQGAQGIL-OH-, tt 581-611), P444 (H-DTQSKNILMQYIKANSKFI-GITF.T.T~KT F~T~T-OH, tt 821-851), P445 (H-IEYNDMFNNFTVSFWLRVPKVS-ASHLEQYGT-OH, tt 941-971), P459 (Ac-VRDIIDDFTNESSQKT-NH2, tt 616-631) and P480 (H-FN~lV~WLRVPKVSASHLE-OH, tt 947-967) were prepared by solid phase peptide synthesis using an Applied Biosystems 430A peptide sythesizer. Peptides were -18148 2180287 PCT/USs3/11703 purifled to >75% and their . - were confirmed by amino acid analysis. TT was a gift from the C .. ~Ith Serum T: ' ' , r ~ ~ , Australia.
PBMC were from l~f~ ;"~ I venous blood of healthy volunteers who had been routinely ' with TT. PBMC were isolated by density-interface . C..o_liUA. over 5 Ficoll-Paque (Pharmacia LKB F ~, 1 Oy AB, Uppsala, Sweden) as described (D.A.
Mutch et aL, Pept Res (1991~ ~L:132-37. The PBMC were . ' ' in complete medium for counting. The average yield of PBMC from whole blood was 2 x 106/mL with a range of 1.2 X l06/mLto 2.9 x 106/mL.
Peptide-stimulated proliferation assays using 2 x 105 PBMC per well were performed in 10 96-well round bottom tissue culture grade mictrotitre plates (Nunc, Roskilde, Denmark).
Antigens were added in 20 ,uL of 0.1 M sodium l,;~ u..~t~ to give a final volume of 200 ,uL per well. Because PBMC often exhibited a low frequency of T cells specific for particular A~ , all assays were carried out using at least 16 replicates per test.
PBMC were incubated at 37C in 5% CO2 Ln humidified air. After 138~t2 h, ~.~.lir~.a~io., was detected by pulsing with 0.5 ~bCi tritiated (methyl-3H) thymidine, 3H-TdR, (40-60 ~ -Ci/mmole, Amersham Australia, Sydney) per well for 6 h. DNA was harvested onto glass fibre filter mats (Skatron, Sterling, VA, USA) and ;~I~UIL ' thymidine was measured in an LKB 1205 Betaplate liquid IIAt;~n counter. All assays included at least 16 wells each of negative controls (20 ,uL of 0.1 M sodium ~ , instead of peptide solution) and positive controls (TT at 1.0 Lf/mL and 0.1 LflmL, also in 0.1 M sodium ~ bu~
buffer) .
The rP~ti~n~lq~ n assay was that of l~.J. Hensen er al., Human Tmmu~ (1984) L:9smodified for use with peptides. For each donor, aliquots of PBMC at 2 x 106/mL in three glass petri dishes were stimulated with TT (1.0 LF/mL), P399 (1.0 ~g/mL) or no antigen respectively. After 6 days, cells were washed to remove any residual peptide or antigen, and ; -A~I in the same volume of fresh complete medium. To test for strength of p~vlirtlAliul. at this time, aliquots of 100 ,uL were pulse-labeled using 3H-TdR. Three days later, further 100 f~L aliquots was tested by the same method to see if ",ulir~ h).l had subsided. The cells from the three petri dishes were then washed and ., .y,. A ~I at a con-centration of 2 x lOs/mL. Replicate aliquots of 100 ~L from each of the three groups were .. . . ..

wo 95/18148 ~ PcrluS93111703 8~
dispensed into wells of U-bottom microtitre trays for 1l ' with the following antigens: P399 at 10, 1.0 and 0.1 ,ug/mL, TT at 1.0 and 0.1 LflmL, or no antigen. To ensure adequate numbers of APC were available, wells were ., ' ' with I x 105 gamma-irradiated (3000 Radl autologous PBMC. After 3 days 1~ (day 12), cells 5 were examined for,ululil`t~liu~ by pulsing with 3H-TdR for 6 hours.
Four regions of the tt sequence were studied. Three of these regions, .,c,ll...,uul.Jil.c tû
peptide pools spanning residues tt 141-171, 581-611 and 821-851 were ! " 1 ' y regions common to the majority of donors tested. T~e fourth region, tt 941-971, 1 ~'(c ,u~ with a reported universally ~ T cell ~Irn-.~ (tt 947-967) was chosen because the 10 pooled 12mers were not ' ,.~ for any of the donors tested. This was therefore a test of the .~ , æ, on peptide length, of stimulation by peptides from a known stimulatûry sequence. PBMC were stimulated with pooled 12mers or with the ~,ull, r ' .E 31mer peptide containing all the residues of the individual peptide pool. Responses were compared using 32 replicates per test to enable drfferences in fi tUU~ S of responses to be 1s ~ c l;~h- .1 The 31mer peptides were tested at a range of ~ " whereas the peptide pool was tested at a single .-..1,~. ..I.~Ii.,-~ comprising 0.3 ~M of each peptide.
Starting from the ~cc.l~rti,~n that the cpnn data for negatives .(llu~ uo~di..c cultures) would be alJUIl ' ' ~y normally distributed about the bd~;hcluulld mean, the lowest values were used for calcuiation of a mean and standard deviation. A temporary cutoff of tll~ mea 20 plus three times the standard deviation was calculated, and all cpm values below this cutof~
were used for calculation of a new temporary cutoff by the salne process. This process was reiterated until no further cpm values were ill~,oluul~lLt d into the r:~ln-l~inn, i.e., until the cutoff remained stable.
After each well was scored as negative or positive, Poisson statistics, which are 25 ~,u~u~u,ul to low frequency events, were used to determine whether any difference between the negative control (cells alone) group and each l .llrl;lll. .l,ll group was significant. Only data significant at the 5 % or better (p < 0.05) level was used.
B. Results Development of the method for T cell d~,t~,. mapping using PBMC

WO9~18148 2~8D287 PCTIUS93/11703 The parameterS of the PBMC IJIVlil'UIdLiOII test were dPtPfminPA using several ~' /h ~101~ protein antigen , and were later retested with a selection of peptide ~' ' found during the course of this work. All the present work used a six day incubation period, chosen because ~lolir~,ldlion usually peaked at that time. In later work, 5 we have opted for a four day incubation to achieve the best ~' between controls and responding cultures.
The need for high numbers of replicates ruled out dose-response titrations on PBMC with each of the large number of peptides used in an initial scan. A target .~ of 0.3~M of each peptide therefore was chosen, based on the limited amount of peptide available 10 and because most T helper cell clones respond to that f,-..ff ~ ;", even though it may not be the optimum. A reported T cell l' t, was recognif ed by the majority of TT-responsive donors. The higher, of peptide (2.4 ,uM) stimulated a higher fref~uency of positive responses from PBMC than the lower cf The use of 0.3 ,uM of a single peptide in a stimulation test could thus lead to failure to detect low frequency 15 responses when they are present. This problem should be lessened when pools of Uv~,lld~- Iq short peptides, offset in start position by one residue, are used.
The choice of 12mers as the peptide length for high-resolution scanning was based on a length of 9 to 11 residues for many known T helper cell .' and on the funding that 12mers were sufficient to detect all major d~L~Illlillall~ of hen egg Iysozyme. For the 20 scanning of large proteins with PBMC, the individual testing of every possible overlapping 12mer becomes completely imprAffifAI, and thus a pooling/decoding strategy was devised. It was found that up to 30 peptides could be pooled for an initial test without ~OIII~JIOIIIi~ill~ the response to individual peptides. Modeling of various ways in which peptides could be grouped revealed that pooling of contiguous overlapping peptides gives the most useful 25 ill~o.l..aii~,n.
The PBMC ,I~ scanning method was developed in studies with tt. Initially, a panel of 12mer peptides IIUIIIOIOCJU~ with the tt sequence as likely T cell .l~t~,l was chosen. One peptide, P399, was found to give proliferation with PBMC of ten out of seventeen TT-responsive donors tested and was therefore chosen for a specificity test.

8~
To see if proliferative responses observed in individual peptide-stimulated PBMC cultures were due to peptide-specific T cells, and to see if these cells were also able to respond to the whole antigen from which the peptide was derived, a ~ ''fjnll assay was carried out. PBMC were stimulated with either peptide P399 (1.0 f g/mL) or TT (1.0 5 Lf/ml). U ' ' control cells were incubated without added antigen under the same conditions. After 6 days in culture, ,UIUlil~,ldLiUI. of antigen-stimulated cultures was - f~ ' ' ' All three groups of cells were washed to remove ly, ' -' and any antigen still present in the medium. On day 9, ululir~ iull had subsided and the PBMC
were washed again. Each culture was then divided into six treatment groups and restim-ulated (P399 at 10, 1.0 and 0.1 ~g/mL, TT at l.0 and 0.l Lflml, and No Antigen control) in triplicate cultures. To ensure that sufficient APC were available, 105 gamma-irradiated autologous PBMC were added to each culture. P.Ulir~ iu.. was measured three days after It' " ' "fm The results fl- ' ' ' that peptide-stimulated PBMC cultures are able to be 1~ ' ' in a dose-dependent fashion by both the sensitizing peptide P399, and the whole antigen TT. Control incubated PBMC not previously stimulated with either antigen showed little ~ùlirt;l~llioll. in the short time frame and with the low number of input unirradiated PBMC, to either the peptide or the whole antigen. This fl' . -1 l l'`'. that peptide-stimulated cultures could be l~,u-~ ly " I ' by the same peptide and were not responding toundefined uullllJ~ of the culture medium. They were also ~ ' ~ by whole TT
showing that they were specific for the antigen from which the peptide was derived. In addition, the TT-sensitized PBMC were 1l I ' by both P39g and TT. Thus, in this instance the cells responding to TT must have included a high proportion of P399-responsive cells also.
To compare the effectiveness of pools of short peptides with that of single long peptides, it is necessary to worli with antigen/donor . ' where specific helper T cells are known to exist. Responses to three dominant ~If ~ regions within the TT sequencefound using peptide pools were compared to responses obtained using 31mer peptides that contain all the residues ~ u ~ f-d by that particular peptide pool. A fourth region 30 spanning a published ''UlUllli~UU.I~'l T helper cell ~' but found to be I ' ,ly , .. ...... .. . .. . ... . ... .. . . .. . . . , . .. . _ _ _ . . . .

~ W095118148 2~0287 Pcr~USs3111703 using pools of peptides, was included to see if a single 3lmer could detect a ~ r~
where the pooling method had failed to do so. The ~ong peptides were tested over a wide range of ~ so that there was no bias against the long peptides due to suboptimal being used. In a pool of 12mer peptides, the effective i.n~ of ' y peptide will depend on the length of the .l. ~ , as those shorter than 12 residues may be IGPI~ ' ' in two, three or more overlapping peptides Although peptide pools 30 and 42 induced IJIUlil~ Liu~ of PBMC at 0.3 ~M/peptide, the ;UII~ -r " ,, 31mer peptides P443 and P444 stimulated few or no responses at any l . ,n.~. 1 ,U ;~ tested. Thus, of the three major ~ . ", ;, ~ regions detected by pools of short peptides, P442 was the only one of the CUIIG~U.. d~l.~ 31mer peptides able to induce ' '- proliferation. Even then, the proliferative responses to P442 occurred at lower frequency than for the peptide pool. The results also d~m~ ~ l that three donors tested were unresponsive to pool 48 or the 31mer peptide P445, which contains the known T cell ,1. ,~ ""j. _..l TT 947-967, despite beimg responsive to the peptide ~;UIIG~ "1" exactly to the published T cell 1~ t, tt947-967. Thus, for this d~tcllllillallL~ neither the longer nor shorter peptides are efficient The most efficient peptide length for specific stimulation of PBMC is unknown. Factors such as the most effficient length for uptake by MHC class lI, small differences in peptide sequences seen by similar T helper clones, and the necessity or otherwise of processing of a 20 peptide before it can be presented will affect the outcome of a stimulation test with pooled peptides on polyclonal T cells (PBMC).
To investigate this problem, four sets of overlapping peptides of different lengths: 10, 12, 14 and 16 residues, spanning a known T helper cell ~ - -containing region of tetanus toxin (INSTKIYSY~ K\TNQGA; tt 587-609), were sy.l~ ~cd. Peptides of each 25 length were then used to make up three separate pools, each pool containing peptides offset by 1, 2 or 3 residues in their l'startl' (i.e., N-terminal) residues. The results (i~ - lU~
that as the offset increases, the frequency of positive responses decreases, as would be expected if not only length but also "frame" of the ~ lll within the peptide were important. For pools of shorter peptides (lû and 12mers) this decrease was so dramatic that 30 for peptides offset by 3, no proliferative responses were observed. This suggests that none WO95118148 i ~ PCT/US93/11703 of the peptides in the pools offset by 3 contained the ! '- ' ' y sequence. In contrast, for 14mer and 16mer peptides, the pools offset by 2 were as effective as those offset by I . In addition, significant responses were obtained using 14mer and 16mer peptides offset by 3.
These results show that as the length of the peptides was increased from 12 to 16 5 residues, the fre4uency of positive responses also increased. A repeat of part of this assay with another sample of the sa[ne lot of cells (frozen PBMC) confirmed these results, although there was a slight variation in test conditions because the peptides had been diluted in complete medium containing 0.2% -Ar~t~nitrilP Other ~ ,, revealed that addition of up to 2% acetonitrile to the culture medium did not affect ~ulir~ iul~ tests on PBMC.
Iû We sought to combine the use of short synthetic peptides, which require little or no processing to be active in T helper cell assays, with the use of unselected PBMC as the source of polyclonal T cells. Using PBMC, the repertoire being examined will not be biased by prior in vitro selection of the best-growing or most frequent clones. We therefore had to solve the combined problem of the limitation in the number of PBMC available from any one 15 donor, and the relatively low frequency of T cells specific to any single l' ' This led us to devise a practical approach involving pools of short peptides and large numbers of replicate cultures.
There are several important limitations when applying such methods to thorough scRening with PBMC. Depending on culture conditions, the most imporiant one being the culture 2û medium, a significant frequency of "false" positives may occur in the llnctimlll ' control cultures. We have minimized the effect of this factor in three ways. Firstly, media other than serum have been thoroughly screened to ensure the lowest ~ k~;luulld " ' while still supporting strong antigen-driven proliferation. Secondly, the number of replicates of ~ controls is made as large as prl~ ti~ The statistical test for 25 a difference between this control group and any test group increases in accuracy with increasing numbers of replicates. Thus, detecting a difference when one is present is more likely, and the estimate of the magnitude of the difference has a higher reliability, as the number of replicates is increased. Thirdly, calculation of a cutoff cpm value to determine the threshold for scoring cultures as positive is based on an method described above. The 3û tests for specificity .1~".,,...:.AI...1 that TT must be processed by these particular donors in wt3 95118148 PCT/US9311l703 ~180~87 such a way that the sequence contained within P399 is generated and bound to MHC class II
molecules for ,u.~ to P399-specific T cells. Not only were TT-sensitized PBMC able to respond to ' by P399, but P39~ d PBMC were also able to react to the whole antigen, TT. Thus, the response of PBMC to the peptide was unlikely to be due S to fortuitous ~IU~Ivd~;~iVi~y with T cells primed by a different antigen.
The test for efficiency of pooled short peptides by ~ , with the ~
31mers showed that the former could be more efficient. This agrees with tests on mouse T
helper clones exposed to ~1t in the context of longer peptides containing added '(~, sequerlces. These results suggest that there can be a block to .vvv~ of 10 long synthetic peptides by helper T cells. There are many reasons why this may occur.
Long peptides may require uptake and processing by specific pathways, compared with smaUer synthetic peptides which can be presented without processing. Antigen processing and ~'t' formation by APC has been shown to vary even for donors with the same restriction element or ., ' mice of identical MHC haplotype, which may be due to a 15 ., , for specific protease(s) for generation of particular ~' The findimgs suggest that within the length range of 12 to 16 residues, longer peptides are more efficient for detecting T helper cell ~' It is not necessary to test pools containing every overlapping peptide along a sequence, if pools containing peptides offset by two residues give an equivalent result. The use of peptides offset by two is advantageous 20 because it s;O.Iir~vdl~y reduces the number of peptides required for systematic synthesis and testing of a given protein sequence. Peptides in the 13 to 18 residue range have been found by extraction from purifled class II molecules, so it appears that synthetic peptides of length similar to the native peptides are the most efficient.
F '- 6 The method of Example 5 can be applied to any antigen of known sequence to whichhuman or animal subjects have a IllVd~llldVle T helper response. We chose to study tetanus - toxin (tt) because it is a commonly used human ~O in which very high r v ~
of specific responding T helper cells occur. (~,~.,.~;,1...,.1,1~ ~'t mapping of tt with 30 human T helper cell clones has been reported, but a limitation of methods used for initia]

, .. ... .

WO95/18148 Q~ PCT/US93111703 c~ .
Iocation of If lf ~ regions was that they relied on efficient processing of protein fragments by pathways similar to those operating with the whole antigen. It has been shown that cells deficient in specific enzymes can fail to process and present a particular peptide despite normal ability to process arld present whole antigen or other peptides.
Using pin technology ~N.J. ~aeji et al., J Immunol Meth (1990) 134:23), we sy.~ cd the set of 1304 overlapping .~ lr sequences spanning the entire 1315 residues of the tt sequence and cleaved them into a l)I,y~i~lo~;i.411y compatible buffer ready for testing.
PBMC from donors shown to respond to tetanus toxoid (TT) in \~itro were screened against peptide pools to locate all ,' in the sequence. Pools were used to keep testing to a ~ lr, realistic scale. PBMC were chosen as the source of responsive T helper cells because they have the advantage of providing a repertoire unbiased by in vitro selection of the best growing or most prevalent clones. When usirlg PBMC, however, large numbers of replicates and controls must be included to ensure statistically significant results are obtained.
Several stimulatory pools were identifled which were common to many donors. T helper cell ~' ~ within such stimulatory pools were then precisely located for several donors by testing the individual peptides comprising each pool. The data thus obtained is more relevant to the total response of human T helper cells to TT than data obtained by ~Li~lLiull of limited portions of the antigen. Three of the five T cell ~Irlr~ I ill -common to the donors tested have not previously been reported.
A set of 1304 overlapping 12mer peptides spanning the tt sequence (U. Eisel et al. ~
EMBO J (1986) 5:2495, il~u~ herein by reference) and offset by one residue was synthesized. The multipin peptide synthesis system, which results in non-toxic peptide solutions ready for use in bioassays, was used. As it was impractical to screen each peptidc separately for its ability to cause ~ tLiOl~ of PBMC, we used a peptide pooling strategy to identify regions along the sequence containing T helper cell ~' We chose to screen peptides as sixty-six pools of ~plu,dlll~t~,ly 20 sequential overlapping peptides each. The size of the pools was selected so that the size of both the initial scan and the subsequent "decodes" of stimulatory pools would be m:ln~r:lhl~. As only those peptide pools stimulating proliferative responses in the first round of testing require "decoding", tlle number of subsequent decodes reduces with increasing number of l~ry pools. It ~ WO9S118148 180~,~7 PCT/US93/11703 should be noted that the peptides from the amino-terminal end of a pool overlap with the pre-ceding pool and likewise the peptides from the carboxy-terminal end of a pool overlap with the following pool.
The c.,"~ of each peptide used im the funal culture was 0.3 ~LM, which was less 5 than the estimated peptide, required to approach opti~mum lqfir n (estimated to be about I ~M). The choice of funal peptide 1~11--- ... lAli~l.. was l~l)~.-ll"i....~ by the need to keep the volume of peptide solution to < lO% of the culture medium to avoid dilution and possible toxicity effects, and also by the peptide ~ ;llll (60 f~M) of the stocksolutions. It should be; ~ ' ' that .' of less than 12 residues in length will 10 be present in 2, 3 or more o.v~ peptides in the pool, and therefore the effective con-centration of shorter d~,l will be higher than that of longer ~
PBMC from nine HLA-typed donors known to respond to TT in virro were irlitially scanned for their ability to respond to each of the 66 peFtide pools. Results are only shown for pools which stimulated a ~ ~ 'y higher number of wells than the cells alone control 15 (p<0.05). Where only one well in a test group showed ~ulir~ iull, even where this was r~ ly higher (p<0.05) than the cells alone control, we chose to treat this as not significant in the sense of .~,.c;,~,..~i,.~ a J~ t., l i '- ~ region of tt. Therefore, such single positive wells were not counted in the summary of donors responding to that pool.
The results showed that many pools stimulated PBMC from more than one donor. Ma~or 20 areas of reactivity, to which more than half of the donors responded, were Pools 30 and 42.
A furfher five pools stimulated four of the nine samples of PBMC. All donors, with the exception of donor "C" responded to a pool unique for that donor while 17/66 (26%) of the pools were not stim~ y for any donor.
The individual peptides within four ' ~r pools were tested to identify the individual 25 peptide(s) I.~u...iblc for proliferative responses incurred by the pool. This test is termed a "decode". Single peptides were tested at I ,uM, dlJ~JIo~dl~ t~.y three times the 1 l ~ 1~"l;,...
of individual peptides used in the pool. The purpose of this was to use a ~on~rntr,qfion close to the effective ,-,.-~ ;.,n which occurs when more than one peptide within a pool contains a ~ ..",i --"l That is, because the length of ~ ..,..i, -"l~ found with T helper cell 30 clones has been eight or nine amino acids, ' Ir sequences of this length would be .. .. . . . .

Wo 95/1814% ~ Pcrluss3lll7o3 present in four or five different overlapping peptides within the pool, making the effective 1.2 ~M or 1.5 ,uM respectively.
The two most common ' y pools, 30 and 42, were decoded. This enabled us to find out whether published T helper cell IL.; could be precisely identified using this S method. Peptides within pool 30 contain sequence YSYFPSVI (tt 593-600), the ~lP~r~
for a human tt-specific T helper cell clone. Pool 42 spans sequence QYIKANSKFIGITEL (tt 830-844), reported to contain a "universally ~ -" DR-restricted d~
Decodimg of pool 30 showed that five overlapping 12mers with start residues 589 to 593 were stimulatory for at least one of the four donors. These 12mers all contain the sequence YSYFPSVI, identical to the published l' In this case. single positive wells wereregarded as m-~ ing~ll because low frequency positive responses appeared to be clustered, i.e., occurred with peptides related by having a high degree of overlap (shared sequence).
Decoding of pool 42 showed eight successive 12mers, with start residues 826 to 833, capable of stimulating PBMC of at least two of the four donors. All these peptides overlap the core of five residues, KANSK, within the reported ~' t, tt 830-844.
As the region tt 579-689 (pools 30 to 34) consisted of five commonly stimulatory pools, we chose to decode two additional pools within this region to identify .1Pl~....;..-..l~ not previously reported Testing of individual peptides within pool 31 and the first two peptides of pool 32 revealed a series of six overlapping stimulatory 12mers witl~ start residues 616 to 620. All these peptides contain the 7mer core sequence IDDFTNE (tt 620-626).
Decoding of pool 33 showed that the response to this pool was due to two distinct . :. . ,.,;..~..~ regions. The T helper cell ~ within this region were centred on sequences IVPYIGPA (tt 642-649) and KQGYEGNFI ~tt 654-662) respectively.
To see if these findings using 12mer peptides would also apply for longer peptides, we 25 synthesized a 16mer which, , ' the "envelope" sequence of the stimulatory peptides from pool 32 (residues 616 to 631; referred to as peptide P459). P459 was tested at two , high (10 ~M or 5 ~M) and low (I ~M), using 8 to 32 replicates per ~.. depending on the number of PBMC available. The results .l~ lAlPd that PBMC from the four donors who had responded to the pooled peptides also responded to P459. Of the 11 donors randomly selected, 10 responded to at least one i5,.~. ~.. I.,.li.,l. of WO95/18148 18~87 pcTilJss3lll7o3 P459, suggesting that these responses were probably not restricted to a single HI~ Class lI
allotype.
To see if a cocktail of dominant ~' of an antigen could stimulate responses as strong or as frequent as responses to the whole antigen, the available ill~l on K
5 l' was utilized. Peptides containing five T helper cell ll. ~ of tt, including ~ previously reported and from the present work, were pooled and the cocktail was tested in parallel with TT.
A ~ Of the rll I of proliferative responses incurred by the two antigen JdldliOl~s usimg PBMC from thirteen donors is shown in. The pool and TT were each 10 tested at four using 8 replicates per . The rl~u~ k" of l;rlldliv~ responses incurred by the pool are generally lower than those incurred by TT.
We have i~ J..~ r~ that stimulation of human PBMC with pools of short syntheticpeptides, followed by decoding to single ' ~Iy peptides, is a practical way to exhaustively map T helper cell l' ~ of entire proteim sequences. Five major T cell 15 :'-' regions along the tt sequence, which appeared not to be restricted to a single MHC class II allotype, were identified. Two of these regions correspond to published T
helper cell .' ~ whereas three have not previously been reported.
The peptide pooling strategy for T cell ~ ..",;,~- ~ i.l. .";ri. ~ has major advantages over the use of protein cleavage fragments or use of long synthetic peptides with small 20 overlaps. All I~V~ J;.Ig peptides of a length within the range of naturally processed peptides (13 to 18 residues) can be ~ ' ' ' without stretching the resources of most research groups. With the poolmg/decoding approach, the task of testing all these short peptides of an antigen on PBMC of individual donors is achievable.
In this work, the choice of 12mers as the peptide length for testing sequences hun~olo~;ou~
25 with tt was based on known minimal ~ and experience with mapping of ~ir~rll..; - -l~ with polyclonal T cells from mice, prior to i~lrolll~d~iu.l on the length of naturally-processed Class II-associated peptides becoming available. Even though the 12mers detected many previously unknown ~Irlrl, ;,~ , had we used longer peptides or more donors we may have detected even more ~1, regions. Thus, the scanning reported 30 here, while more thorough than any previously reported, probably does not represent the .. . . . . . .

WO 95/18148 ~ PCT/US93/11703 total spectrum of d.,, for tt because it does not include donors of all MHC types tested at all peptide lengths. While carrying out the tt ~ r~ scanning reported here using every o;~ arl ~ 12mer, we found that pools of slightly longer overlapping peptides (14 to 16 residues in length) offset by two or three residues are also effective. The number S of peptides needed can thereby be . ~ , reduced, although there is usually a small reduction in sensitivity of detection of peptide-responsive cells when peptides offset by more than one residue are used. A reduction in sensitivity in these ~ may be due to adecrease in the effective ~.. . - :.,,li.. of ' y sequences (since each .;. ~.. ;. -.. 1 is ' in fewer peptides), or due to non-optimal N- and C-terminal residues in the 10 peptides used.
The "decoding" of adjacent peptide pools showing significant stimulation of PBMC can distinguish between two ~)n~ s The ctir~ Iy sequence(s) in successive positive pools of overlapping peptides could occur at the boundaries of the pools. Both pools may then contain the same stimulatory sequence, despite the impression that two .1..~ ....;.--..1~ are 15 involved (donors B and D responding to pools 31 and 32). Alternatively, donors responding to adjacent pools may be responding to ;~ , unrelated .~..~...,..; --.1~ separated by y sequences (donor B responding to pools 30 and 31).
Within the four pools decoded, there were cases where at least four overlapping 12mer peptides were stimulatory. This may be due to a minimal ~.t~ ll of 9 residues shared 20 among all four 12mers, but because the PBMC system is polyclonal, these peptides could b~
activating the progeny of more than one T cell clone. The proliferative response to related peptides could thus be due to activation of clonal progeny of one precursor T cell by a sequence common to the peptides, or due to activation of a number of i 1~ l~ 1. . T cell clones able to respond to similar (but different) sequences within this region. A test of these 25 alternatives would be to use peptides of the minimum length needed to bind to MHC Class n molecules and stimulate T helper cells. This would enable the effect of small differences in recognised sequences on the measured frequency of specific clonal progeny in PBMC to b~
seen.
Contrary to first d~ r~ . the physical length of our nominal " 12mer" stimulatory 30 peptides is consistent with the 13 to 18 residue length range of the peptides bolmd to Class II

WO95118148 802&~7 PCT/uss3/11703 antigens. This is because our peptides all consist effectively of 15 residues, having 12 residues of the tt sequence with a constant tripeptide moiety (b-dkp) at the carboxy-terminal end. We have found that identical 12mer peptide sequences lacking the b-dkp group (i.e., with free carboxy-termini) are less effective at stimulating PBMC. An acetylated amino ter-S minus can also lead to increased effectiveness of T helper (If lf . i ~ peptides. T helpercell clones can be stimulated by b-dhp bearing peptides of 8, 9 or 10 residues, suggesting that peptides containing 12 residues of the antigen sequence have more than the required amount of sequence needed to allow MHC class n binding and le~u~ by the helper Tcell receptor.
The finding that the amino terminus of the peptide is an important and consistent part of the peptide that binds to MHC Class II antigens suggests that peptides differing in amino-terminal position by only one residue would activate different pop~ nc of T helper cells. If this is the case, then testing smaller numbers of longer peptides could result in failure to detect some ~l, t~ , since peptides with the required N-terminal residues may 15 not be present in the pool. We have found that a series of contiguous overlapping peptides from within ' y pools are l y, suggesting that the N-terminal residue is not critical in flft..,..i.~- mapping with PBMC.
APC play a critical role in antigen-stimulated PBMC proliferation assays. Short synthetic peptides can be efficiently presented by a range of APC, including B cells, monocytes, 20 dendritic cells and possibly other APC. It is likely that emcient uptake and ~ ,Af~,l,dli.n. of 12mer peptides by APC in PBMC occurs in our c;~ ~ I conditions, because addition of adherent cells from autologous PBMC does not cause an increase in the frequency of positive responses. It is known that short peptides can be taken up directly by MHC class II mol-ecules without being processed, but the relative ~ ",- ri~ ~e of this pathway versus an 25 i. ~ , pathway for peptides interacting with APC in PBMC is unknown at this time.
For longer peptides, however, inefficient detection of precursor T cells may be occurring, since certain pools of 12mers were stimulatory for PBMC in contrast to 31mer peptides spanning the same sequences as the stimulatory pools.
One valuable ~ of the present work is that it shows how all the T helper cell 30 .I~,t~,l of an antigen could be identified by an unbiased method, leading to a greater WO 9Stl8148 PCTtUS93tll703 o~8~
' ~ of many aspects of the immune system, including the basis of ,' selection and the factors in peptide sequences underlying MHC Cl~ss II restriction. This may enable accurate prediction of T helper cell 1' from primary sequence data alone.
S E~nowledge of the T cell 1l ' ' ' of an antigen will allow design of reagents for or . 1 Peptides containing T cell ~I~ .f. .. ~ may be able to be used alone or in ..~ i.... with whole antigens to increase the ~ / of vaccines. A peptide which helps the formation of IgG rather than IgE may alleviate an allergy. Alternatively, a d~.t~ ,o..~;bl~ for an: disease could be 10 modified to lead to tolerance and therefore alleviation of the disease.
Pools of T cell ,1 may also be used as effective substitutes for whole antigens in diagnostic ~uliLIdLiu~l assays. Testing of the pool of peptides containing five human tetanus toxin T helper cell 1' ' shûwed that all donors tested recognized at least one within the pool. This pool of peptides thus represents a l~;lJIUd.~ T cell 15 " I y antigen mixture which may help the ' di~Liul. of T cell ~)IUIi~ iUli tests worldwide.

Claims (14)

WHAT IS CLAIMED:

1. A T cell epitope peptide selected from the group consisting of , and .

2. A T cell epitope peptide comprising a sequence of at least 8 consecutive residues selected from a sequence selected from the group consisting of , and .

3. The T cell epitope peptide of claim 2, comprising twelve residues selected from a sequence selected from the group consisting of , and .

4. The T cell epitope peptide of claim 2, further comprising a detectable label.

5. A method for detecting exposure of a mammal or bird to an antigen, which method comprises:
obtaining a biological sample from said subject, said sample containing T cells;contacting said T cells with a T cell epitope peptide; and detecting reaction between said T cell and said T cell epitope peptide.

6. The method of claim 5, wherein said T cell epitope peptide comprises a sequence of at least 8 consecutive residues selected from a sequence selected from the group consisting of , and .

7. The method of claim 6, wherein said T cells are contacted with a plurality of T cell epitope peptides.

8. An assay kit for detecting exposure of a mammal or bird to an antigen, which kit comprises:
a T cell epitope peptide comprising a sequence of at least 8 consecutive residues selected from a sequence selected from the group consisting of , and ; and printed instructions for performing said assay.

9. The kit of claim 8, wherein said T cell epitope peptide further comprises a detectable label.

10. A vaccine composition for inducing an immune response in a bird or mammal, said composition comprising:
an effective amount of a B cell antigen;
an effective amount of a T cell epitope peptide comprising a sequence of at least 8 con-secutive residues selected from a sequence selected from the group consisting of , and ; and a pharmaceutically acceptable carrier.

11. The vaccine composition of claim 10 wherein said B cell antigen comprises HSV gD2 or HIVSF2 gp120.

12. A method for increasing the number of T cells in a subject specific for a selected antigen, which method comprises:
obtaining a biological sample from said subject, said sample containing T cells;contacting said T cells with a T cell epitope peptide;
culturing said T cells to specifically expand T cells reactive to said T cell epitope peptide:
and administering G said T cells to said subject.

13. A method for determining T cell epitopes specific to an antigen, which method comprises:
preparing a plurality of peptide pools, wherein each pool comprises a plurality of peptides, wherein each peptide comprises eight amino acids, wherein the sequence of each peptide is selected from the sequence of the antigen;
obtaining a biological sample from a bird or mammal, said sample containing peripheral blood mononuclear cells;
contacting said peripheral blood mononuclear cells with said peptide pools;
culturing said peripheral blood mononuclear cells in the absence of xenogeneic serum;
and determining the mitogenic effect of said peptide pools.

14. The method of claim 13, wherein each peptide comprises twelve amino acids, and wherein each successive peptide overlaps the sequence of the preceding peptide by at least eight amino acids.