AU651346B2 - Method for measuring T-cell surface antigens in humans - Google Patents

Description

METHOD FOR MEASURING T-CELL SURFACE

ANTIGENS IN HUMANS

RELATED APPLICATION

This application is a continuation-in-part of U.S. Patent Application Serial No. 437,370, filed November 15, 1989.

FIELD OF THE INVENTION

This invention relates to a method for diagnosing a pathological condition, via assaying or measuring

particular T-cell subtypes in a sample taken from a patient suspected of having the pathological condition. In particular, it relates to measuring cell surface antigens of T-cells which are characteristic of

particular T-cell subtypes.

RELATED PUBLICATION

Portions of the invention described herein have been presented in Kappler, et al., Science 244: 811-813 (May 19, 1989), the inventors' publication and the disclosure of which is incorporated by reference herein. BACKGROUND AND PRIOR ART

In recent years, the mechanism by which mammalian immune systems, such as human and murine systems react to infections, foreign antigens, and to so-called "self antigens" in connection with autoimmune diseases has begun to be established. See, in this regard. Grey, et al., Scientific American 261 (5): 56-64 (1989); Male, et al., Advanced Immunology (J.P. Lippincott Company, 1987), especially chapters 6 through 10. Well known, both to the skilled artisan and to the general public is the role of antibodies, sometimes referred to as "immunoglobulin" or the less correct and older "gammaglobulin" in response to infection.

Antibodies are protein molecules which are produced by B cells in response to infection. It is well known that these antibodies act to "disable" or to inactivate infectious agents in the course of combating the

infection.

In order for antibodies to be produced, however, preceding events must occur which lead to stimulation of the B cells which produce the antibodies. One of the key events involved in the processes leading to antibody production is that of antigen recognition. This aspect of the immune response requires the participation of so-called "T-cells", and is less well known than the antibody response commented on supra.

Briefly, and in outline form, antigen recognition requires interaction of an "antigen presentation cell", a "processed antigen", and a T-cell. See Grey and Male, supra. The "processed antigen", in an infection, is a molecule characteristic of the pathogen which has been treated, i.e., "processed", by other cells which are a part of the immune system. The processed antigen

interacts with a receptor on the surface of an antigen presented in a manner not unlike a lock fitting into a key hole or, perhaps more aptly, two pieces of a jigsaw puzzle.

The configuration of the complex of processed antigen and receptor on antigen presentation cell allows the participation of T-cells. T-cells do not join the complex unless and until the processed antigen has fit into the receptor on the antigen presentation cell. This receptor will hereafter be referred to by its scientific name, the major histocompatibility complex (MHC), or the human leukocyte antigen (HLA). Generally, MHC is used to refer to murine systems, and HLA to humans. These receptors fall into two classes. MHC-II molecules are involved in most responses to pathogens. In contrast, MHC-I molecules are involved when the pathogen is a virus, or a malignant cell is involved. When MHC-I participation is involved, there is no antibody stimulation; rather, the interaction of MHC-I, processed antigen and T-cell leads to lysis of cells infected with the pathogen.

The foregoing discussion has focused on the events involved in responding to "infection", i.e., the presence of pathogenic foreign material in the organism. Similar mechanisms are involved in autoimmune diseases as well. In these conditions, the organism treats its own

molecules as foreign, or as "self-antigens". The same type of complexing occurs as described supra, with an antibody response being mounted against the organism itself. Among the diseases in which this is a factor are rheumatoid arthritis, diabetes, systemic lupus

erythromatosus, and others.

The ability of the T-cell to complex with the processed antigen and MHC/HLA complex is dependent on what is referred to as the T-cell antigen receptor, referred to as "TCR" hereafter. The TCR is recognized as a heterodimer, made up of alpha (α ) and beta (β ) chains. Five variable elements, coded for by germline DNA and known as "Vα, Jα, vβ, Dβ, and Jβ " as well as non-germline encoded amino acids contribute to the TCR. See, in this regard, Marrack, et al., Immunol. Today 9 : 308-315 (1988); Toyonaga, et al., Ann. Rev. Immunol 5: 585-620 (1987); Davis, Ann. Rev. Immunol 4: 529-591

(1985); Hendrick, et al., Cell 30: 141-152 (1982). With respect to the binding of TCR with processed antigen and MHC, see Babbitt, et al., Nature 317: 359-361 (1985);

Buus, et al., Science 235: 1353-1358 (1987); Townsend, et al., Cell 44: 959-968 (1986); Bjorkman, et al., Nature 329: 506-512 (1987). Generally, both the alpha and beta subunits are involved in recognition of the ligand formed by processed antigen and MHC/HLA molecule. This is not always the case, however, and it has been found that so-called

"superantigens" stimulate T-cells with a particular V/β element, regardless of any other element. See Kappler, et al., Cell 49: 273-280 (1987); Kappler, et al., Cell 49: 263-271 (1987); MacDonald, et al., Nature 332: 40-45 (1988); Pullen, et al., Nature 335: 796-801 (1988);

Kappler, et al., Nature 332: 35-40 (1988); Abe, et al.,

J. Immunol 140: 4132-4138 (1988); White, et al., Cell 56: 27-35 (1989); Janeway, et al., Immunol. Rev. 107: 61-88 (1989); Berkoff, et al., J. Immunol 139: 3189-3194

(1988), and Kappler, et al., Science 244: 811-813 (1989). This last reference discloses information which is also incorporated into the subject patent application.

The "superantigens" mentioned supra, while generally stimulating T-cells as long as they possess a Vβ

element, are somewhat specific in terms of the particular form of the Vβ moiety which is present on the

stimulated T cell. This feature is one aspect of the invention, i.e., the ability to assay for particular subtypes or subclasses of T-cells, based upon the cell surface antigens presented by these subclasses.

Staphylococcus aureus has long been implicated in morbidity and mortality in humans. See Bergdoll, in Feed Bourne Infections and Intoxications (Riemann and Bryan, ed., Acad. Press, N.Y.) pp. 443-494 (1979). The various toxins presented by S. aureus are responsible for most food poisoning cases, as well as severe shock, and other life threatening pathological conditions. The mechanism of action of the toxins associated with S. aureus is unknown. The primary structure of the toxins, while showing some relationship, also show some great

differences in primary structure. See Betley, et al., J. Bacteriol 170: 34-41 (1988); Jones, et al., J. Bacteriol 166: 29-33 (1986); Lee, et al., J. Bacteriol 170: 2954-2960 (1988); Blomster-Hautamaa, et al., J. Biol. Chem. 261: 15783-15786 (1986). For the time being, it cannot be said with any certainty whether the various S. aureus antigens function in the same way in terms of the immunological response they generate.

The ability of S. aureus to stimulate powerful T cell proliferative responses in the presence of mouse cells bearing MHC-II type molecules is taught by, e.g., Carlson, et al. J. Immunol 140-2848 (1988); White, et al., Cell 56 27-35 (1989); Janeway, et al., Immunol. Rev. 107: 61-88 (1989). White, et al., and Janeway, et al. showed that one of these proteins is not mitogenic, in that it selectively stimulates murine cells which bear particular Vβ elements. These papers, however, did not extend the study to human cells. It has now been shown, however, that certain antigens do selectively stimulate specific Vβ subclasses of human T cells, making it possible to diagnose pathological conditions by assaying for particular Vβ subtypes.

Hence, it is an object of the invention to describe a method for diagnosing a pathological condition in a human by assaying a biological sample from the subject being tested for levels of particular Vβ subtypes.

These levels are then compared to normal levels, where a difference between the two is indicative of a

pathological condition.

It is a further object of the invention to carry out the assaying using antibodies which are specific for the particular Vβ subtype. Especially preferred are

monoclonal antibodies.

It is still another object of the invention to perform the above described assay by measuring DNA coding for specific Vβ molecules. This can be done via

utilizing, e.g., the polymerase chain reaction.

How these and other objects of the invention are achieved are detailed in the disclosure which follows. BRIEF DESCRIPTION OF THE FIGURES

Figure 1 shows the results of staphylococcal toxin stimulation of human T cells.

Figure 2 depicts studies showing Vβ specific stimulation of T cells by toxins is donor independent.

Figure 3 depicts a standard curve used to normalize polymerase chain reaction values (PCRs) to percentages of T cells carrying particular Vβs in mixed populations.

Figure 4 shows autoradiograms of coamplified cDNA of human TCR transcripts following stimulation with anti-CD3 antibody or a S. aureus toxin.

Figure 5 presents in bar graph form Vβ specific

stimulation caused by S. aureus toxins in three

individuals.

Figure 6 shows autoradiograms of T cell receptor

transcripts amplified by polymerase chain reaction from cells Patient 1 (P) and control individual (C).

Figure 7 shows longitudinal changes in T cell repertoire in 2 patients studied serially after toxic shock

syndrome.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

Example 1

This experiment used monoclonal antibodies directed against Vβ5, vβ6 , Vβ8 and Vβ12, as taught by Yssel, et al., Eur. J. Immunol. 16: 1187 (1986); Borst, et al., J. Immunol. 139: 1952 (1987); Posnett, et al., Proc. Natl. Acad. Sci. USA 83: 7888 (1986); Carrel, et al., Eur. J. Immunol 16: 649 (1986) , and Bigler, et al., J. Exp. Med. 158: 1000 (1983).

T cells of a human individual were first isolated from that individual's peripheral blood. These T cells were then examined before and after stimulation with one of (i) anti-CD3 antibody, (ii) SEC2; (iii) SED, or (iv) SEE. Items (ii), (iii) and (iv) are known S. aureus molecules which act as toxins.

The anti-CD3 antibodies had been rendered

stimulatory by adherence to plastic bottles. The protein was incubated on plastic surfaces for 8 hours at 4°C. Extensive washing removed non-adherent antibody.

Following this, either adherent antibody or a S. aureus antigen was used to stimulate peripheral blood T cells.

Stimulation took place in the presence of

irradiated, autologous, non T-cells as described by

Kotzin, et al., J. Immunol. 127: 931 (1981), the

disclosure of which is incorporated by reference herein.

Three days after stimulation, live cells were collected and cultured for 24 hours in recombinant human IL-2 (25 units/ml). This allows regeneration of

potentially modified receptors. Of the surviving cells, about 10% were true blast cells.

The blast cell fractions were then incubated with one of (i) purified antibody to CD3 or with a monoclonal antibody to (ii) Vβ5 (mAb 1C1); (iii) Vβ6 (mAb OT145); (iv) Vβ8 (mAb MX6), or (v) Vβ12 (mAb S511). Following incubation with the mAb, the cells were stained with fluoroscein-conjugated goat anti mouse IgG, following Kappler, et al., Cell 49: 173 (1987). The staining pattern was then studied on an EPICS C device, using a forward angle and 90° light scatter pattern to gate large blast cells, which were easily distinguished from small lymphocytes, and constituted 50% or more of all surviving cells in culture.

The results of the staining patterns are shown in Figure 1. Panels A-D shows the degree of staining using the mAbs before stimulation. Panels E-H show it after stimulation with anti-CD3. Finally, panels I-L show the pattern following stimulation with SED, SEE, and SEC2.

Each anti-Vβ stained a definable percentage of the peripheral resting T cells from this donor (Fig. 1). The percentage stained ranged from 5.2% with anti-Vβ6 to 1.5% with anti Vβ12 (Fig. 1, A to D). Culture with anti-CD3 and interleukin-2 hardly changed the percentage stained with each anti-Vβ (Fig. 1, E to H), indicating that this combination of T cell stimuli affected T cells bearing different αβ receptors similarly. Culture with the toxins had variable effects on the percentage of T cells stained with each anti-V/6 (Fig. 1, I to L).

Staphylococcal entertoxin (SE) D, for example, greatly increased the percentage of T cells bearing Vβ5 in the blast population and nearly excluded cells bearing Vβ6. In contrast, T cells blasts stimulated with SEC2 were depleted of Vβ6- and Vβ8 bearing T cells and were greatly enriched in Vβ12 bearing T cells. Finally, SEE

stimulated Vβ8 T cells, while excluding cells bearing Vβ12. Reciprocal results for each of the toxins were found if the resulting T cells contaminating the blast populations were analyzed for Vβ usage. After SEE stimulation, for example, the resting T cells were selectively depleted of Vβ8⁺ cells. This result

indicates that the toxins are stimulating most of the T cells bearing the appropriate Vβs, nor a minor population of these cells.

Five different donors were used in the experiments. These donors were HLA-typed by standard serological techniques, and their restring peripheral T cells were stained with anti-CD3 and the anti-Vβs. Each of the anti-VAs reacted with a low but measurable percentage of peripheral blood T cells from each of the individuals (Table 1). For a particular individual these percentages were extremely reproducible from one day to another. The percentages of T cells that bore the different Vβs varied somewhat among individuals.

Example 2

Cells from the different donors were stimulated with anti-CD3 or the staphylococcal toxins and analyzed for CD3 and Vβ expression (Fig. 2). For each individual, results were calculated as the percentage of T cell blasts bearing a particular Vβ after stimulation divided by the percentage of T cells bearing that Vβ before stimulation. This calculation was designed to correct for variations in Vβ expression from one person to another. As before, anti-CD3 stimulated T cells bearing the different Vβs uniformly; the ratio of T cells bearing a particular Vβ before and after CD3 stimulation was close to 1. In contrast, it was clear that the staphylococcal toxins varied markedly in their ability to stimulate T cells bearing different Vβs. For example, T cells bearing Vβ5 and Vβ12 were quite rich in blasts produced by challenge with SEC3 , whereas T cells bearing Vβ8 were specifically excluded from the SEC3 blasts.

One or more of the toxins was a stimulus for T cells positive for each of the Vβ families (albeit weakly for Vβ6) , indicating that a toxin superantigen had been identified for each of the Vβ families. Conversely, toxins could be identified which specifically failed to stimulate T cells bearing each of the Vβs.

It is remarkable that a characteristic stimulation pattern could be identified for almost each toxin. SEC2, for example, stimulated T cells bearing Vβ 12 and

excluded cells bearing Vβ s from the other three

families. This pattern was not seen with any of the other toxins. SED stimulated T cells bearing Vβ5 and Vβ 12, had marginal effects on T cells bearing Vβ8, and excluded cells bearing Vβ6. Again, this pattern was unique to this toxin.

In some cases, stimulation with a given enterotoxin yielded blasts that were neither enriched nor depleted for expression of a given Vβ by comparison with the starting population. Starting and ending percentages of Vβ 5-bearing cells were similar, for example, in

responses to toxic shock toxin (TSST). Such a result might indicate that only some Vβ5-bearing T cells were stimulated by TSST. Perhaps the other variable

components of the receptor, Vα , Jα , or Jβ, could quite often prevent interaction of this toxin with Vβ5, a phenomenon that has been noticed before for superantigen reaction with mouse T cell receptor Vβs. Alternatively, TSST may react with only one member of the Vβ5 family. Thus, in responses to TSST, the increase in blasts bearing this member may be offset by a disappearance of T cells bearing other members of the family, but also reactive with 1C1. Discrimination by superantigens among different members of Vβ families has been seen in mice, where the self superantigen Mls-1 stimulates T cells positive for Vβ8.1 but not those bearing Vβ8.2 or

Vβ8. 3 (Kappler, et al. Nature 332: 35 (1988), and SEC1 stimulates T cells bearing Vβ8.2 but not those bearing Vβ8.1 or Vβ8.3.

In some experiments, the percentages of T cells that stained with anti-CD4 or anti CD9 were checked before and after stimulation. The starting percentages were virtually unchanged by toxin stimulation. T cells from one donor, for example, were initially 78% CD4⁺ and 23% CD8⁺. After stimulation with the nine different toxins the percentages in the blast of CD4⁺ cells ranged from 74% to 79%, and of CD8 cells from 20% to 25%, suggesting that all these stimuli affected CD4 and CD8 cells

equally. It might have been expected that the toxins, which are dependent on class II MHC for presentation would have preferentially stimulated CD4⁺ cells, but such is not the case.

One of the most striking features of the data in

Fig. 2 is the consistency of the results from one

individual to another. Thus, although the five people tested had different HLA types and different starting percentages of T cells bearing the various Vβs (Table

1), the proportional changes in Vβ expression in blasts stimulated by each toxin were almost the same from one individual to another. Although the superantigens require class II MHC for presentation, the allele of class II has much less impact on superantigen

presentation than it does on recognition of conventional antigens plus MHC by T cells.

These results show that the staphylococcal toxins are not indiscriminate mitogens for human T cells, but are, in fact, Vβ-specific. This result accounts for the previously noted clonal specificity for such toxins.

Although each toxin is able to stimulate only a

subpopulation of all T cells in humans, they are still powerful T cell stimulants, active at low concentrations. Some or all of the toxic effects of these proteins in humans may be mediated by their ability to stimulate large numbers of human T cells. For example, the ability of these toxins to induce secretion of large quantities of lymphokines is probably secondary to their ability to stimulate, in a vβ - specific way, a sizable percentage of T cells. It is also possible that the ability of these and other microbial-derived superantigens to stimulate populations of T cells bearing particular Vβs may be related to the differential resistance of

different individuals to the effects of these toxins and also to the ability of microbial attack to induce immune consequences, such as autoimmunity, in certain

individuals.

Example 3

The foregoing examples demonstrated a method for quantifying T cell subsets having particular cell surface phenotypes, using antibodies. This methodology calls for interaction between the antibody and its binding

partners, i.e., the cell surface antigen, which is the Vβ molecule.

Enhanced presence of the VA molecules means that there has been enhanced expression of the DNA coding for the particular molecule. Thus, the following experiments deal with the measurement of the aforementioned T-cell subsets via analysis of the DNA expressing a particular Vβ subtype.

Among the methods available to the skilled artisan for analyzing DNA is the so-called polymerase chain reaction, or "PCR" as used hereafter. PCR methodology is well known to the art, as may be seen in, e.e.g, U.S.

Patent Nos. 4,683,195, 4,683,202 and 4,800,159, Saiki, et al., Science 239: 487-491 (1988), and Chelly, et al., Nature 333: 858-860 (1988). Given that the PCR

methodology is known to the art, only the modifications to the technology used are elaborated upon.

Total RNA was prepared from anti-CD3 stimulated peripheral T cells as described supra. Two μg of total

RNA was used for the synthesis of first strand cDNA using reverse transcriptase (Amersham) and random

hexanucleotides. The reaction was stopped by heating for 5 minutes at 95°C before polymerase chain reaction.

One twentieth of each cDNA samples was co-amplified using a V β-specific primer with a Cβ primer and two Cα primers as set forth at Table 2 with final concentration of 0.3 μM in each reaction. The amplification was performed with 2.5 U of Taq polymerase (Perkin-Elmer) and a Cetus Perkin Elmer thermocycler under the following conditions; 95°C melting, 55°C annealing, and 72°C extension for 1 minute each. For quantification of amplified products, coamplification was performed with 5' 3² P-labelled reverse primers (about 5x10⁵ cpm each). The amplified products were separated on 2% agarose gels, dried and exposed to X-ray film. The autoradiograms were used to identify and cut out the Vβ -Cβ and Cα bands.

Each band was counted by liquid scintilation counter. In control experiments, the relative amplification

efficiency was calculated essentially as described by

Chelly et al., supra.

Notes To Table 2

The size of amplified products (Vβ bands) by Vβ and 3'Cβ primers ranged from about 170 to 220 bp. The size of the amplified cDNA (Cα band) by 5'Cα and 3'Cα primers was about 600 bp. The 3'Cβ primer used in this study matches exactly both Cβ1 and Cβ2 DNA. The sequences of Vβ , Cβ , and Cα are from previously published reports.

aMembers of each Vβ family which have identical sequences as the corresponding primer are listed.

b Vβ 13.1, ^cVβ13.2, and ^dVβ14.1 have also been called Vβ 12.3, 12.4, and 3.3, by Toyonaga, et al., Ann. Rev. Immunol. 5: 585-620 (1987), Kimura, et al., Eur. J. Immunol. 17: 375-383 (1987).

Among the at least 20 different families of human Vβ genes, at least 46 different members of these families have been cloned and sequenced, as reported by Toyonaga, et al., supra; Concannon, et al., Proc. Natl. Acad. Sci. USA 83: 6598-6602 (1986); Lai, et al., Nature 331:

543-546 (1988). To analyze human T cell Vβ usage, 22 different Vβ-specific oligonucleotides for use as 5' sense primers for PCR were synthesized. Their sequences, and the Vβ's which they would be expected to amplify, are shown in Table 2. All the Vβ's indicated as

amplified have sequences matching their corresponding primers exactly. There may have been other Vβ genes amplified with these primers. For example, the Vβ6 primer matches Vβ6.4 except for one nucleotide, and further experiments will be needed to find out if Vβ6.4 is amplified using this primer. Altogether, all these primers would be expected to cover at least 39 of the 46 sequenced human genes. Each Vβ specific oligomer was picked to have roughly the same G+C content and to be located at relatively the same position in Vβ. Example 4

Total RNA was prepared from human peripheral T cells stimulated by anti-CD3 antibody or one of 5 different S. aureus toxins (SEB, SEC2, SEE, exfoliating toxin

(ExT), and toxic shock syndrome toxin 1 (TSST), as described in the previous examples. At the time of analysis these populations contained 50-90% T cell blasts as judged by flow cytometric analysis. A single strand complementary DNA was prepared for mRNA phenotyping, following Buus, et al., Science 235: 1353-1358 (1987); Townsend, et al., Cell 44: 959-968 (1986), and aliquots of cDNA from each sample were amplified with each of the 22 5' Vβ specific sense primers and the 3' Cβ specific antisense primer. As an internal control, TCRα chain mRNA was co-amplified in the same tube. Amplification was performed with 25 cycles, a limited number used to ensure that the amount of product synthesized was

proportional to the amount of Vβ mRNA in the original preparation. The specificity of each Vβ specific primer was determined by the size of its amplified product and hybridization to the amplified products of specific probes (not shown). The amplification efficiencies of four of the primer sets (5'Cα-3'α , Vβs 2 , 3, and 8-3' Cβ) were determined as described by Chelly, et al., supra. The average efficiency ranged about 46-48% . For each sample the number of counts in the Vβ band were

normalized to those found in the Cα band.

It was necessary to find out whether or not the relative incorporation in this PCR reaction was

proportional to the number of cells in the responding population expressing a particular Vβ element. However, two possible sources of error had to be considered. The first of these was contribution from unstimulated T cells. It was reasoned that, since mRNA levels are extremely low in unstimulated T cells compared to T cell blasts, the contribution from unstimulated cells would only become a problem when the proportion of blasts expressing a particular Vβ was very low compared to the unstimulated cells. Secondly, since all T cells have the potential to rearrange the β-locus on both chromosomes, transcription of Vβ mRNA from a non-productively

rearranged chromosome in at least some T cells might confuse the analysis. Since non-functional mRNA could be expected to be at a low level due to its instability, it was reasoned that this mRNA may only present a problem in cases where a particular Vβ element was poorly expressed in the blast population.

In order to test these assumptions, the actual percentage of T cell blasts expressing Vβ 5.2/3, Vβ 8 and Vβ12 in the various samples using flow cytometry and anti- Vβ monoclonal antibodies was determined prior to preparing mRNA. When the normalized PCR incorporations for Vβ's 5.2/3, 8 and 12 for these samples were plotted in a log/log plot against the percentage of T cell blasts staining with these anti-Vβs monoclonal antibodies, a linear relationship was obtained (Figure 3) with the data from three different experiments indistinguishable. This relationship was most evident for values above 1%. Below about 1% Vβ expression or a normalized PCR incorporation of about 30 the correlation was lost. It was concluded, therefore that contributions from unstimulated T cells and non-productively rearranged β-genes were

insignificant when Vβ expression in the blasts was greater than 1%. Therefore, the data plotted in Figure 4 was used as a standard curve to analyze expression of Vβ's for which antibody was not used, estimating the percent VA expression from the normalized PCR

incorporation.

Example 5

The PCR methodology was used to analyze the

expression of Vβ 5.2/3, Vβ8, Vβ12 and 19 other Vβs or Vβ families in normal peripheral T cells stimulated with the various toxins. T cells stimulated with anti-CD3 were used as a control, because Examples 1 and 2 show that stimulation with anti-CD3 did not significantly change the percentages of T cells bearing particular Vβ's from that seen in the starting population. Results are shown in Figure 4. The results of a complete

analysis of the response of T cells from a single

individual to five different S. aureus toxins are

summarized in Table 3.

Some Vβ families were used more abundantly than others by normal peripheral T cells. Members of the Vβ 2 , 3, 6, 7 and 8 families and Vβ13.1 were expressed by more than 50% of total T cells. Such a finding was perhaps not unexpected for VA6 and VA8 which are part of large families of Vβ's (although the Vβ6

oligonucleotide probably primes for only 3 of the 9 members of the Vβ6 family), but is more surprising for Vβ13.1, which appears to be the product of a single gene. The uneven expression of Vβ's by human peripheral T cells did not appear to be idiosyncratic for this individual or determined by MHC, since similar frequencies were seen for 2 other unrelated human donors tested (see discussion, infra, and Figure 5).

Complete analysis of the expression of mRNA for all 20 families of human T cell receptor Vβ genes showed clearly that all the toxins preferentially stimulated T cells expressing particular Vβ 's, moreover the pattern of stimulation was different for each toxin. A number of striking new associations were found. Most dramatically Vβ2-bearing cells were highly-enriched by stimulation with TSST. About 50% of the T cells in TSST stimulated T cell blasts had Vβ2. As was shown, supra SEB stimulated T cells bearing Vβ12, but this analysis also revealed stimulation of T cells bearing Vβ3, Vβ14, Vβ15, Vβ17 and perhaps Vβ 20 by SEB. The related toxin, SEC2, also stimulated T cells expressing Vβ12, Vβ14, Vβ15, Vβ17 and Vβ20, but not those expressing Vβ 3 . SEE stimulated T cells bearing members of the Vβ8 family, as we have previously shown, but also increased the proportion of Vβ5.1⁺, Vβ6.1-3⁺, and Vβ18⁺ cells.

Using this method, it was possible to estimate roughly the percentage of all the T cells in a given human cell population that could be accounted for by summing those bearing the different Vβs measured. As shown in Table 3, this percentage was about 90% for T cells stimulated with anti-CD3, suggesting that the estimate that the Vβ oligonucleotides would prime for expansion of mRNA's encoded by 39 of the 46 human Vβ genes is not exaggerated, certainly not by an order of magnitude. This suggests that the 46 known Vβ sequences probably cover most of the human genes. The quantitative

PCR's accounted for a lower percentage of blasts

stimulated by some of the toxins, in particular, ExF. It is possible that this toxin predominantly stimulates T cells bearing Vβ's not covered by the listed primers.

Some of the most dramatic associations in Table 3 were tested in two additional human individuals to see how general the phenomena were (Figure 5). The

stimulation experiments, and calculations, were identical to those used supra. In their responses to these toxins the 3 individuals behaved almost identically. For example, Vβ 2⁺ T cells were enriched by TSST to almost the same level of 45% in every case. Similarly, in all three individuals, SEB stimulated T cells bearing Vβ 3 and SEE stimulated T cells bearing Vβ 8 . Example 6

The similarities between mice and humans in the T cell response to these toxins in striking. In both cases T cells bearing particular Vβ's dominate the response to each toxin. In both cases the discriminatory powers of the toxins can be particularly dramatic. For example, in humans Vβ5.1 T cells responded to SEE, whereas cells bearing Vβ5.2/3 did not. Similarly, it has been

observed by the inventors that, in the mouse, several toxins can distinguish among the members of the Vβ8 family. This member-specific response to superantigens has also been seen in mice for the endogeneous

superantigen, Mls-1 , which stimulates T cells bearing Vβ8.1 but not those expressing V β 8 .2 or Vβ8.3. See Kappler, et al., Nature 332; 35-40 (1988).

Extensive sequence analysis of Vβ genes from mouse and man shows that there are some homologues, both by primary sequence, and by their relative location in the Vβ gene complex. See Toyonaga, et al., Ann. Rev. Immunol. 5: 585-620 (1987); Concannon, et al., Proc. Natl. Acad. Sci. USA 83: 6598-6602 (1986); Lai, et al., Nature 331: 543-546 (1988). The stimulation patterns by the

different toxins of these homologues by using data for mice Vβ stimulation by toxins was compared, following White, et al., Cell 56: 27-35 (1989). As indicated in Table 4, in some cases T cells bearing homologous Vβ's show a similar pattern of response to the toxins. ExT and especially TSST, for example, stimulated T cells bearing human Vβ2 and mouse T cells bearing the most analogous Vβ 15. Human T cells expressing members of the Vβ12, 14, 15 and 17 families all showed a tendency to respond to SEB and SEC2, but not ExT or TSST. This property was shared by their closest murine relatives , mouse V β ' s 8 .1 , 8 .2 and 8.3. However, similar response patterns by T cells bearing homologous Vβ's was not always seen. For example, T cells bearing murine Vβ3 responded to most of these toxins, however, those bearing the closest human analog, Vβ10, did not. Even with all this information in hand, a close examination of the primary amino acid sequences of the human and mouse Vβ elements has not yet revealed the essential residues responsible for toxin specificity. Thus, while tempting, complete generalization from mouse to human systems (MHC to HLA) is not indicated.

In comparing mouse and man, the most striking difference to emerge thus far in our studies is the apparent lack of mechanisms limiting Vβ expression in humans. In the mouse, despite the potential for

expression of over 20 Vβ elements in the species as a whole, various mechanisms limit Vβ expression in

individual mice. In some strains large genetic deletions have eliminated about half of the Vβ gene elements.

See, e.g., Behlke, et al., Proc. Natl. Acad. Sci. USA 83; 767-771 (1986). Other Vβ gene elements are often inactivated by point mutations. See Wade, et al., J.

Immunol. 141; 2165-2167 (1988). Most ingeniously, in many strains of mice, self-superantigens, expressed during T cell development lead to the deletion of T cells bearing particular Vβ elements during the establishment of self tolerance. See in this regard Kappler, et al., Cell 49: 273-280 (1987); Kappler, et all, Cell 49:

263-271 (1987); macDonald, et al., Nature 332; 40-45 (1988); Pullen, et al., Nature 335; 796-801 (1988);

Kappler, et al., Nature 332; 35-40 (1988); Abe, et al., J. Immunol 140; 4132-4138 (1988). It is proposed that these mechanisms which lead to limited Vβ expression in individual mice may be a protective evolutionary response to the pressure exerted by bacterial toxins, so that in a population of mice some individuals will be relatively resistant to the effects of any particular toxin

superantigen. No evidence for widespread similar

mechanisms in humans has emerged thus far from the limited number of individuals examined. Thus large genetic deletions have not been found nor have

self-superantigens which cause elimination of T cells bearing particular Vβ been observed. A closer

examination both of individual members of the Vβ

families and of larger human populations, especially those with a much more widespread exposure at an early age to these types of toxins, may be required to observe some of these mechanisms at work in humans. Example 7

Patients (9 in total) were all diagnosed as having toxic shock syndrome by their private physicians. They were then screened to determine if they met the

definition for severe toxic syndrome as jointly created by the Centers for Disease Control and several

investigators. See Todd, Clin. Microbiol. Rev. 1:

432-466 (1988); Reingold, et al., Ann. Intern. Med. 96 :

875-80 (1982); Wisenthal, et al., Ann. J. Epidemol. 122; 847-56 (1985). Major criterial (all required) for the diagnosis include fever (≥39.8°C), rash (diffuse

erythematous rash evolving to desquamation), and

hypotension (systolic blood pressure <90 mmHg for adults and/or orthostatic syncope or dizziness). Minor criteria

(3 required) for the definition include diarrhea and/or vomiting, muscular involvement, mucous membrane

hyperemia, decreased renal function or pyuria, elevated liver enzymes, platelet count <100,000/mm³, and

disorientation or altered state of consciousness. All patients studied also had at least one probable focus of

S. aureus infection. As a control, in addition to normal individuals, one patient with severe toxic shock syndrome associated with group A Streptoccocus pyogenes (Stevens, et al., New Eng. J. Med. 321; 1-7 (1989) was also

studied.

Disease in four patients appeared to be related to menstruation and tampon use. S. aureus was cultured from

2 of these patients, and the vagina was presumed to be the site of S. aureus infection in the remainder. The development of toxic shock syndrome in Patient 6 appeared to be related to an S . aureus vaginal infection four weeks after cesarian section. Disease in Patient 4 was associated with sinusitis from which S . aureus was cultured . Eposides in two children (Patients 7 and 8) were associated with subcutaneous abcess of the buttocks and peritonsillar cellulitis, respectively. S. aureus was cultured from both foci, and these isolates produced TSST-1 as well as entertoxins A and C in vitro. Patient 3 had previously experienced more than 10 episodes of toxic shock syndrome thought to be related to upper respiratory and sinus infection with S. aureus. Despite being prophylactically treated with dicloxacillin, she required hospitalization for the clinical episode of toxic shock syndrome studied here. At the time of hospitalization, a culture of the nasopharynx was

negative for S. aureus. Thus, with the possible

exception of this latter patient, toxic shock syndrome in the nine patients studied appeared to be related to focus of S. aureus infection. The time from acute onset of symptoms and from initiation of treatment (including antibiotics) to the time samples were obtained for analysis of T cell subset changes is also listed in Table 5. At the beginning of the study, it was unclear as to how long abnormalities in T cell subsets would persist after the toxic shock syndrome , and no restriction was placed on this variable.

^*Major criteria include fever (>102°F), a characteristic erythematous rash that subsequently is followed by

desquamation, and hypotension.

Minor criteria include A, gastrointestinal symptoms

(vomiting or diarrhea); B, muscle involvement (cleaved

CPK or sever myalgias); C, hyperenia of the mucous

membranes; D, renal functional impairment or pyuria; E, laboratory evidence of hepatic dysfunction; F,

thrombocytopenia (platelet count <100,00/mm³); and G, disorientation or altered state of consciousness. #Patient 3 had at least 10 prior episodes of toxic shock syndrome. Several of these had been associated with

sinusitis, and previous evaluations had included positive cultures of S. aureus from the nasopharynx and sinus. The patient was being treated prophylactically with antibiotics, and at the time of this hospitalization, all cultures were negative for S. aureus. +⁺ In some of the later cases studied, the source of

infection was cultured and S. aureus isolates were tested for in vitro production of enterotoxins, A,B,C₁,C₂,C₃,D and E, and TSST-1. ^§ Not done.

Peripheral blood mononuclear cells taken from the patients were isolated from heparinized blood following Ficoll hypaque gradient centrifugation. Stimulation of T cells were accomplished as described supra in plastic flasks coated with purified anti-CD3 antibody. Cells were cultured at 1x10⁶ cells per ml in media containing RPMI 1640 supplemented with 2 mM glutamine, 10 mM hepes buffer, 100 u/ml penicillin, 100 ug/ml streptomycin, and 10% fetal calf serum. After three days of culture, live cells were harvested and cultured for an additional 24 hours in 25 μ/ml recombinant human IL-2 to expand cells expressing the receptor for IL-2 while allowing

regeneration of potentially modulated T cell receptors. The cells were then harvested, washed and used for either indirect immunofluorescence staining or quickly frozen in liquid nitrogen for subsequent RNA extraction.

Indirect immunofluorescence was performed by

incubating cells with saturating amounts of monoclonal antibody, and then staining with a fluorescein-conjugated goat anti-mouse Ig(Tago, Inc., Burlingame, CA) was described supra. Control samples included cells stained with the second-step reagent alone, and background values were subtracted. Forward angle and 90°C light scatter patterns were used to gate on the large lymphoblasts, which were easily distinguished from small lymphocytes and which comprised the great majority of all viable cells. Fluorescence intensity was determined using an Epics C cell sorter. Monoclonal antibodies used as staining reagents were directed to CD4 and CD8, and to epitopes on T cell receptors bearing Vβ5, VA6, Vβ8, and Vβ12, again as described supra.

Total RNA was prepared from anti-CD3 stimulated cells as described and total RNA was used for the synthesis of first strand cDNA using reverse

transcriptase, also as indicated supra. Similarly, cDNA was amplified using the polymerase chain reaction as described in Example 3. In a fashion similar to Examples 1-6, peripheral blood mononuclear cells were isolated from patients and controls, and T cells were stimulated in culture with anti-CD3 antibody and IL-2. Using cDNA made from total RNA isolated from the T cell blasts and a quantitative polymerase chain reaction, T cell receptor gene segments encoding Vβ2 , Vβ5 (5.2 and 5.3), Vβ8 (8.1 and 8.2), and Vβ12 were amplified and quantitated. To control for the amount of T cell receptor mRNA and variation in reaction rate, a Cα gene segment was also amplified in each reaction. Figure 6 shows results with T cells from Patient 1 and a concomitantly-studied normal individual. A striking increase in amplified Cβ2 DNA is apparent in the patient compared with control, whereas little

difference is observed in the Cα products. There is also little difference between patient and control in the amount of Vβ5, Vβ8, or Vβ12 product especially when normalized to the relative amount of Cα amplified in each reaction.

The data in Table 6 are expressed as a ratio of Vβ DNA to Cα DNA amplified in the same reaction. Although all of the controls (studied concomitantly with patients) had Vβ2/Cα ratios less than or equal to 0.10, initial samples from 5 of the 9 patients had ratios greater than 0.17 (p=0.03 by Fisher's exact test) with one, for

Patient 1, as high as 0.78. Abnormal Vβ2/Cα ratios were demonstrated in 3 of the 4 patients with

menstruation-related disease and in 2 of the 5

nonmenstrual cases. In contrast to Vβ2, none of the other Vβ2/Cα ratios were increased more than 2 SD above control values, indicating the selective nature of

Vβ2 expansion in toxic shock syndrome. It should be noted that among the patients not demonstrating increased Vβ2/Cα ratios (Patients 4,6,8,9), one (Patient 4) was studied a relatively long period after the acute disease such that an increased level could have been missed.

Controls (N=7)

(mean±S.E) 0.09±0.01 0.05±0.01 0.025+0.01 0.03±0.01

Data arc presented for the initial sample obtained from each patient (see Table 1). + A primer specific for a sequence common to Vβ5.2 and Vβ5.3 family members was used

# A primer specific for a sequence common to Vβ8.1 and Vβ8.2 family members was used.

T cells isolated from patients and controls were also analyzed for subset alterations by indirect

immunofluoresce and cytofluorographic analysis. No consistent alteration m the percentage of CD4⁺ and CD8⁺

T cells were noted even in patients with markedly

elevated Vβ2 levels. The ranges of CD4 and CD8⁺ T cell percentages were 30-84% and 14-60%, respectively, in patients compared with means (± SE) values of 64 ± 5.1% and 29 ± 4.6% for control individuals. Cells from 7 of the patients were also stained for the expression of Vβ

5 , Vβ6, Vβ8, and Vβ12. Consistent with the above analysis by the quantitative polymerase chain reaction technique, values outside the normal range were not observed for these non-Vβ12 T cell subsets. Mean ± SE values for patients vs. controls (N=12) were: Vβ5,

2.4±0.25 vs. 3.1±0.30; Vβ8, 4.3±0.63 vs. 3.7±0.50; Vβ 6,

3.0±0.46 vs. 3.4±0.41; and Vβ12, 1.6±0.14 vs. 1.5±0.09.

Examples 1-6 show that the efficiency of Vβ2 amplification is similar to that for Vβ5 and Vβ8, supporting the validity of estimating the percentage of circulating T cells expressing Vβ2 by the polymerase chain reaction method. The results suggest that Vβ2⁺ T cells in normal individuals are approximately 10% of the peripheral blood T cell population. In contrast, peak values for patients 1 and 2 may be as great as 70% and

30%, respectively, emphasizing the striking stimulation of Vβ2⁺ T cells occurring in some patients.

One patient with severe toxic shock syndrome

associated with group A streptococcus was also studied. All major and minor criteria for the definition of toxic shock syndrome were present, and the patient died

approximately too weeks after hospitalization. The patient was studied within one week of the onset of symptoms, and the Vβ2/Cα ratio was 0.08, clearly within the normal range. Example 8

Serial samples were obtained from two patients to examine longitudinal changes in T cells expressing Vβ2 after the acute disease. Vβ2 T cell percentages in Patient 1 decreased by half within 2 weeks after the initial determination and were almost normal by 60 days after the onset of the acute disease (Figure 7 - top panel). Patient 2 demonstrated near-normalization of Vβ2 levels within 45 days of the acute episode (Figure 7 - bottom panel). These serial studies also emphasize the relative lack of fluctuation over time in T cell subsets expressing other Vβ segments, an observation also made in studies of normal individuals some of which are discussed supra.

The foregoing examples show that a pathological condition, such as an infection, can be diagnosed by assaying a sample from a patient to determine levels of particular Vβ molecules in the sample. Increased levels of specific subtypes have been found to be linked to particular antigens, as the results show.

The term "pathological condition" as used herein is not limited to an infection; rather, it refers to any condition where an abnormal immune response has occurred. This includes, e.g., autoimmune diseases where, as has been shown supra, specific Vβ type molecules are present where they should not be, or are present in quantities above those found in normal individuals.

Increases in Vβ quantities are not the only way to diagnose pathological conditions in accordance with this invention. The art is familiar with various diseases and pathological states, such as HIV infection, where T cell levels are below those normally encountered. Correlation of particular Vβ types to conditions characterized by T cell depletion are also embraced herein.

Examples 7 and 8 shows that a selective increase in circulating T cells expressing Vβ2 is frequently

associated with toxic shock syndrome. Other T cell subsets studied including those expressing Vβ5, Vβ8, and Vβ12 were not increased above normal levels and did not fluctuate over time. Thus, measurement of the proportion of T cells expressing Vβ2 in peripheral blood could be used as a diagnostic test for this disease.

Diseases that result in nonspecific T cell activation should not lead to a selective increase in Vβ2⁺ T cells.

It should also be emphasized in TSST-1 and other S.

aureus enterotoxins stimulate T cells bearing particular Vβ segments almost regardless of the composition of the rest of the T cell receptor on these cells. Other variable elements (Dβ , Jβ, Vα , Jα) do not appear to contribute much to the recognition of these Vβ-specific superantigens as they do for conventional antigens.

Thus, although it is possible that some conventional antigens may also stimulate a subset of T cells expressing

Vβ2, the magnitude of this response with the frequency of responding T cells being much less than 1 in 100 would be only a small fraction of that resulting from

stimulation of TSST-1. It seems unlikely that such a response could change circulating levels of Vβ2⁺ T cells.

It is apparent that not all of the patients with toxic shock syndrome in this study demonstrated elevated

Vβ2 levels. One patient was studied nearly five months after the acute illness, and elevated levels could have been missed. This possibility is corroborated by the serial changes in two patients, in which a return to near-normal levels occurred within one to two months after presentation. Examples 1-6 indicated that S.

aureus toxins other than TSST-1 do not stimulate Vβ 2⁺ T cells but do stimulate other sets of T cells in a vβ- specific fashion. Thus the experiments set forth in

Examples 7 and 8, which focused on T cells expressing

Vβ2 , are likely to detect abnormalities only in TSST-1 mediated disease. Future studies that include the

measurement of T cell subsets likely to be expanded by these other toxins may increase the frequency of positive tests. As predicted, the patient with fatal shock

syndrome associated with group A streptococcus did not have elevated levels of circulating Vβ2⁺ T cells.

The data presented here indicate that during toxic shock syndrome T cell stimulation occurs on a scale not observed in response to conventional antigens, and it is proposed that this massive T cell activation is a critical event in the development of disease. These activated T cells are likely to be releasing IL-2, interferon-gamma, lymphotoxin (TNF-β), and a variety of other less

well-characterized lymphokines. See Micusan, et al., Immunology 58: 203-8 (1986). IL-2 infusions have been associated with a high frequency of induced hypotension as taught by Belldegrun, et al., Ann. Int. Med. 106:

817-22 (1987); Ann. Int. Med. 109: 953-8 (1988). The T cell activation process and/or release of IL-2 could also greatly enhance or be required for the release of IL-1 and TNF by macrophases. Nedwin, et al., J. Immunol. 135: 2492-7 (1985). If T cell activation is required for expression of toxic shock syndrome, T cell depletion or functional inactivation should interrupt the cascade of events in S. aureus toxin mediated disease. This may be partially supported by studies suggesting that

administration of corticosteroids early in the disease course may produce beneficial effects in some patients. See Todd, et al., JAMA 252: 3399-3402 (1984).

In this initial study, patients were not studied at the time of initial presentation. Kinetic studies where therefore limited to following changes in T cell

repertoire after the initial increase in percentage of Vβ2⁺ T cells. The return of this subset to relatively normal levels was surprisingly rapid in the two patients that were serially followed. The fact that levels do return to normal indicates that the increase in Vβ2⁺ T cells occurs after the onset of toxic shock syndrome and is not a factor influencing susceptibility. The mechanisms responsible for the decrease in circulating Vβ2⁺ T cells with time are unclear. Studies in rodents after immunization have indicated that after activation and expansion in numbers, antigen-specific T cells emigrate from lymphoid tissues (the site of antigen activation) into the recirculating pool (i.e. thoracic duct lymph). See Sprent, et al., Cell Immunol. 2: 171-81 (1971); Wilson, et al., J. Immunol. 116: 1030-40 (1976). These cells gradually decrease over time, perhaps

reflecting their return to lymphoid tissues and/or the continuous entry of other antigen-activated cells into the recirculating pool.

The terms and expressions which have been employed are used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, it being

recognized that various modifications are possible within the scope of the invention.

Claims (10)

WE CLAIM:

1. Method for diagnosing a pathological condition in a human comprising assaying a body fluid sample of a patient to determine level of a T-cell subtype

characteristic of said pathological condition and

comparing the level determined to normal levels in a comparable body fluid sample, wherein a variance in the level of said T-cell subtype is indicative of said pathological condition.

2. Method of claim 1, wherein said pathological condition is a staphylococcus infection.

3. Method of claim 1, wherein said pathological condition is an autoimmune disease.

4. Method of claim 1, comprising assaying said body fluid sample by contacting it with an antibody which specifically binds to a cell surface antigen

characteristic of said T-cell subtype.

5. Method of claim 4, wherein said antibody is a monoclonal antibody.

6. Method of claim 5, wherein said monoclonal antibody is a monoclonal antibody specific for a T-cell surface antigen selected from the group consisting of Vβ 5, Vβ 6 , V β 8 and Vβ12.

7. Method of claim 1, comprising assaying said body fluid sample by determining the amount of DNA present which codes for a cell surface antigen which

characterizes said T-cell subtype.

8. Method of claim 7, comprising determining said DNA by polymerase chain reaction.

9. Method of claim 7, wherein said cell surface antigen is selected from the group consisting of Vβ1, Vβ2, Vβ 3, Vβ4, Vβ 5.1, Vβ 5.2, Vβ 5.3, Vβ 6.1, Vβ 6.2, Vβ6.3, Vβ7, Vβ8, Vβ9, Vβ 10, V β11, V β12, Vβ13.1, Vβ13.2, Vβ14, Vβ15, Vβ16, Vβ17, Vβ18, Vβ19 and Vβ20.

10. Method for diagnosing toxic shock syndrome comprising assaying a body fluid sample of a patient to determine level of Vβ2 subtype T-cells in said body fluid sample and comparing said level to normal levels in a comparable body fluid sample, wherein an increase in said Vβ2 subtype level is indicative of toxic shock syndrome.