The Diversity of Antigen-Specific TCR Repertoires Reflects the Relative Complexity of Epitopes Recognized

Antigen-selected T cell receptor (TCR) repertoires vary in complexity from very limited to extremely diverse. We have previously characterized two different CD8 T cell responses, which are restricted by the same mouse major histocompatibility complex (MHC) class I molecule, H-2 K^d. The TCR repertoire in the response against a determinant from Plasmodium berghei circumsporozoite protein (PbCS; region 252–260) is very diverse, whereas TCRs expressed by clones specific for a determinant in region 170–179 of HLA-CW3 (human) MHC class I molecule show relatively limited structural diversity. We had already demonstrated that cytolytic T lymphocyte (CTL) clones specific for the PbCS peptide display diverse patterns of antigen recognition when tested with a series of single Ala-substituted PbCS peptides or mutant H-2 K^d molecules. We now show that CW3-specific CTL clones display much less diverse patterns of recognition. Our earlier functional studies with synthetic peptide variants suggested that the optimal peptides recognized were 9 (or 8) residues long for PbCS and 10 residues long for CW3. We now present more direct evidence that the natural CW3 ligand is indeed a 10-mer. Our functional data together with molecular modeling suggest that the limited TCR repertoire selected during the CW3 response is not due to a paucity of available epitopes displayed at the surface of the CW3 peptide/K^d complex. We discuss other factors, such as the expression of similar self MHC peptide sequences, that might be involved in trimming this TCR repertoire.     

The epitope detected by cytotoxic T lymphocytes against thymus leukemia (TL) antigen is TAP independent

Thymus leukemia (TL) antigens belong to the family of MHC class Ib antigens. We have shown in our previous studies that they serve as transplantation antigens, and can be recognized by both TCR alpha beta and TCR gamma delta cytotoxic T lymphocytes (CTL) with TL but not H-2 restriction. Although TL are known to be expressed TAP independently, it is unclear whether peptide loading on TL molecules is necessary for the formation of CTL epitopes. In the present study, we first showed that TL expression is beta(2)-microglobulin (beta(2)m)-dependent but TAP1 independent by flow cytometric analysis of thymocytes from beta(2)m- or TAP1-deficient mice crossed with TL transgenic mice expressing Tla(a)-3-TL on their thymocytes. Subsequently, we investigated the epitope recognized by CTL derived from C3H mice immunized with skin from a transgenic mouse expressing T3(b)-TL ubiquitously. Bulk CTL lines against TL from primary mixed lymphocyte cultures showed comparable cytotoxicity against T3(b)-TL transfectants of TAP2-deficient murine RMA-S grown at 37 degrees C to that against those grown at 25 degrees C. Furthermore, TCR alpha beta and TCR gamma delta CTL clones against TL recognized TL expressed on T3(b)-TL transfectants of RMA-S and Drosophila melanogaster cells having broad defects in peptide loading of MHC, and lysed these target cells. These results together indicate that TL-specific CTL populations primarily recognize epitopes expressed TAP independently.     

Restricted alpha/beta receptor gene usage of idiotype-specific major histocompatibility complex-restricted T cells: selection for CDR3-related sequences.

We have sequenced the T cell receptor (TcR) V alpha and V beta genes of seven independent BALB/c CD4+ T cell clones specific for the immunoglobulin lambda 2 light chain produced by the MOPC 315 myeloma (lambda 2(315)). All the clones recognize a peptide of residues 91-101 of lambda 2(315) and are restricted by the major histocompatibility complex (MHC) molecule I-E(d). The results indicate that in BALB/c mice, this anti-idiotypic response uses a very limited number of TcR. The four clones which cross-react between Phe94 and Tyr94 peptide analogues use very similar receptors (V alpha 3, J alpha 1, V beta 6, J beta 1.1). The V alpha 3 gene used by all of these clones is identical and has not been previously described. Although the four clones differ in nucleotide sequence in the V/J borders, two had identical receptors at the amino acid level. One of the cross-reactive clones exhibits a heteroclitic response to the Tyr94 peptide variant resulting from a single amino acid exchange in the V/J junction of the alpha chain. The three remaining clones which recognize only the Phe94 and not the Tyr94 peptide have somewhat more diverse TcR, however, two of these three clones use V beta 6. One of these non-crossreacting clones is alloreactive, the specificity of which can be attributed to differences in the N-D-J sequences. Taken together these data indicate that this T cell response to an immunoglobulin idiotope is very restricted in terms of the TcR used. These data in conjunction with recently published results indicate that, although there can be strong preference for individual V alpha or V beta gene segments, certain V alpha/V beta combinations are preferentially selected for interacting with a given peptide/MHC combination, and that the CDR3-related regions are crucial for antigen fine specificity and alloreactivity.     

Evidence for monoclonal expansion of synovial T cells bearing V alpha 2.1/V beta 5.5 gene segments and recognizing a synthetic peptide that shares homology with a number of putative autoantigens.

A peptide of 15 amino acids derived from the cereal glycine-rich cell wall protein (GRP), sharing a significant homology with Epstein-Barr virus nuclear antigen-1 (EBNA-1), fibrillar and procollagen, stimulated synovial fluid (SF) T cells from juvenile (JRA) and adult (RA) rheumatoid arthritis patients. An overexpression of the V alpha 2 gene family was found in the SF from patients who responded significantly to the peptide. To investigate in more detail the SF T-cell responses to the GRP peptide, we established peptide-specific T-cell lines and clones from a DR8+ positive JRA patient with pauciarticular form. The T-cell clones were phenotyped as T-cell receptor (TCR)alpha beta+/CD4+ and their clonality was investigated by polymerase chain reaction (PCR) and flow cytometric analysis. TCR sequences from different clones demonstrated that the clones were identical and used the V alpha 2.1/J alpha 6 combined with V beta 5.5/J beta 2.7 gene segments. Interestingly, direct sequencing of the V alpha 2 family PCR product obtained from cDNA prepared from freshly isolated SF mononuclear cells identified the same TCR sequence as that used by the clones, suggesting the monoclonality of SF CD4+ T cells bearing V alpha 2.1/J alpha 6 gene products. The present data suggest a recruitment and expansion of a SF T-cell subpopulation, and also support the hypothesis that autoimmune diseases can be triggered by protein epitopes with crucial amino acids homologous to self-proteins.     

Antitumor cytotoxic T-lymphocyte response in human lung carcinoma: identification of a tumor-associated antigen

Summary: We have isolated several cytotoxic T lymphocyte (CTL) clones from lymphocytes infiltrating a lung carcinoma of a patient with long survival. These clones showed a CD3⁺, CD8⁺, CD4^–, CD28^– phenotype and expressed a T‐cell receptor (TCR) encoded either by Vβ8‐Jβ1.5 or Vβ22‐Jβ1.4 rearrangements. Functional studies indicated that these clones mediated a high human leukocyte antigen (HLA)‐A2.1‐restricted cytotoxic activity against the autologous tumor cell line. Interestingly, TCRβ chain gene usage indicated that CTL clones identified in vitro were selectively expandedin vivo at the tumor site as compared to autologous peripheral blood lymphocytes (PBL). These findings provide evidence that an immune response may take place in non‐small cell lung carcinoma and that effector T cells may contribute to tumor regression. Further study indicated that the CTL clones recognized the same decamer peptide encoded by a mutated α‐actinin‐4 gene. Using tetramers of soluble HLA‐A2 molecules loaded with the mutated antigenic peptide, we have derived several anti‐α‐actinin‐4 T‐cell clones from patient PBL. These CTL, recognizing a truly tumor‐specific antigen, may play a role in the clinical evolution of this lung cancer patient. Adoptive transfer of CTL clones in a SCID/NOD mice model transplanted with autologous tumor supported their antitumor effect in vivo.     

Evidence for shared recognition of a peptide ligand by a diverse panel of non-obese diabetic mice-derived,islet-specific,diabetogenic T cell clones

MHC class II-restricted autoreactive T cells play a major role in the development of autoimmune diabetes mellitus in both human and mouse. Two of our groups previously established panels of islet-reactive CD4+ T cell clones from prediabetic non-obese diabetic (NOD) mice. These clones express distinct sets of TCR V alpha , V beta , J alpha and J beta , and also differ in the structure of the junctional region of TCR. All of the T cell clones have been shown to cause insulitis and several induce diabetes when transferred to various recipients. The antigen specificities of these T cell clones have not been determined, but they do not react with defined islet cell antigens such as glutamic acid decarboxylase. To identify the peptide ligands recognized by these clones, we examined the reactivity of the T cell clones to peptide mixtures in which anchor residues for H2-A g7 were fixed. Most of the clones showed similar reactivity to the peptide mixtures. To further determine the peptide ligands of the T cell clones, we synthesized several peptides based on the favored amino acid motifs and examined clone reactivity to the synthetic peptides. Some of the peptides, e.g. HLAI-RM and HIPI-RM, could stimulate most of the T cell clones tested, even though the clones expressed different TCR. The results suggest that our islet-reactive T cell clones recognize in islet beta cells a natural ligand that is similar to these peptides.     

Qa-1/Tla region alloantigen-specific CTL with alpha beta receptor

Recent studies involving T cells that express gamma delta T-cell receptor (gamma delta TcR) have raised the possibility that Qa-1/Tla region class I major histocompatibility complex (MHC)-like molecules are antigen-presenting molecules for gamma delta TcR. In this report, cytotoxic T lymphocyte (CTL) clones specific for a Qa-1/Tla region gene product were isolated from a bulk B10. QBR (Kb, Ib, Dq Qa-1/Tlab) anti-B10.MBR (Kb, Ik, Dq, Qa/Tlaa) CTL line. These CTL lysed blasts from all Qa-1a strains regardless of the H-2 haplotype, indicating that the recognition of the Qa-1 antigen by these CTL is not restricted by other class I molecules. In bulk populations, CTL activity of this specificity was found only in the CD8+CD4- subpopulation. Accordingly, all established CTL clones were phenotyped as Thy-1+, CD8+CD4-. Furthermore, these clones were shown to express alpha beta TcR rather than gamma delta TcR. Thus, the results indicate that Qa-1 antigen can be recognized by alpha beta TcR T cells in a manner similar to recognition of classical class I molecules.     

Unconventional cytotoxic T lymphocyte recognition of synthetic peptides corresponding to residues 1—23 of Ras protein

A peptide corresponding to amino acids 1 through 23 of Ras protein containing a mutation at position 12 was used to induce cytotoxic T lymphocytes (CTL) in mice. Although the CTL were CD8⁺ and expressed α, β T cell antigen receptors (TCR), their major histocompatibility complex (MHC)-restriction was unconventional. They recognized peptide-treated murine cells of different H-2 haplotypes, but not MHC class I-negative cells. Human HLA class I molecules did not present Ras peptides and hybrid human/mouse MHC molecules revealed that all three extracellular domains α1, α2 and α3 were required for recognition by peptide-specific CTL. Shortening the 23-mer peptide by 5 residues at either the amino or carboxy terminus resulted in loss of CTL recognition. This demonstrates an unusual form of antigen recognition by mouse CTL in which peptide presentation requires murine H-2 class I molecules but is not class I allele restricted, and the peptides recognized are much larger than peptides in conventional class I-restricted responses.     

CD8alpha/alpha homodimers fail to function as co-receptor for a CD8-dependent TCR

In this study, we have started to dissect the molecular basis of CD8 dependence of a high and low avidity CTL clone specific for the same peptide epitope. Using anti-CD8alpha and anti-CD8beta antibodies, we found that cytotoxicity and IFN-gamma production by high but not by low avidity CTL was strongly CD8 dependent. We isolated the TCR genes of both types of CTL clones and used retroviral gene transfer to analyse the function of these TCR in primary T cells of wild-type and CD8beta-deficient mice. Both TCR triggered antigen-specific killing in wild-type T cells, and blocking experiments showed that CD8 dependence/independence co-transferred with the TCR into primary T cells, indicating that it was dictated by the TCR itself. Gene transfer experiments into CD8beta-deficient T cells revealed that only the TCR derived from the CD8-independent CTL clone elicited antigen-specific cytotoxicity, while the CD8-dependent TCR was non-functional in the absence of the CD8beta-chain. These data indicate a striking difference between CD8alpha/beta heterodimers and CD8alpha/alpha homodimers as only the former were able to provide co-receptor function for the CD8-dependent TCR.     

A single CTL clone can recognize a naturally processed HIV-1 epitope presented by two different HLA class I molecules

Although it is known that a single peptide can be recognized by CTL restricted to two MHC class I alleles, there is no direct evidence for presentation of a single peptide by two MHC class I molecules. Furthermore, it is unclear whether such peptides are presented to the same T cell or to different T cells. Our previous study suggested that CTL recognition of the human immunodeficiency virus-1 (HIV-1) Pol HIV-B35-SF2-24 epitope (IPLTEEAEL) occurs via both HLA-B35 and HLA-B51 restriction. Here we provide the first direct evidence that a single CTL clone can recognize this peptide presented by both HLA-B35 and HLA-B51. Furthermore, we directly purified this peptide eluted from both HLA-B*3501 and HLA-B*5101 molecules isolated from target cells infected with HIV-1 recombinant vaccinia virus. These results demonstrate that HIV-B35-SF2-24 is a naturally processed peptide which is presented by both HLA-B*3501 and HLA-B*5101. TCR analysis of one CTL clone suggested that it is a single clone. B*3501-SF2-24-tetrameric complexes inhibited both HLA-B*3501- and HLA-B*5101-restricted recognition of this clone, suggesting that the TCR of this clone cross-recognize the structure of both HLA class I-peptide complexes.     

Functional differences between influenza A-specific cytotoxic T lymphocyte clones expressing dominant and subdominant TCR.

We have shown that the dominance of CD8+ T cells expressing TCR Vbeta17 in the adult HLA-A*0201-restricted influenza A/M1(58-66)-specific response is acquired following first antigen exposure. Despite the acquired dominance of Vbeta17+ cells, subdominant M1(58-66)-specific clones expressing non-Vbeta17+ TCR persist in all individuals. To determine whether the affinity of the expressed TCR for the HLA-A*0201/M1(58-66) complex could influence functional properties, M1(58-66)-specific clones expressing subdominant (non-Vbeta17+) TCR were compared to cytotoxic T lymphocyte (CTL) clones expressing dominant (Vbeta17+) TCR. The Vbeta17+ CTL required up to 10,000-fold lower amounts of M1 peptide to mediate lysis compared to CTL clones expressing other Vbeta gene segments. All Vbeta17+ CTL clones tested bound HLA-A*0201/M1(58-66) tetramer, but two of three CTL clones expressing other TCR did not bind tetramer. The inability of non-Vbeta17+ CTL to bind tetramer did not correlate with phenotype, CD8 dependence or with cytokine production profiles. This suggests a limitation for the use of tetramers in examining subdominant T cell responses. Together these findings suggest that Vbeta17+ CTL which dominate the HLA-A*0201-restricted CTL response against influenza A are not functionally distinct from subdominant non-Vbeta17+ CTL. The dominance of Vbeta17+ CTL is likely to result from a competitive advantage due to superior CTL avidity for the HLA-A*0201/M1(58-66) complex.     

Role of alpha3 domain of class I MHC molecules in the activation of high- and low-avidity CD8+ CTLs

CD8 can serve as a co-receptor or accessory molecule on the surface of CTL. As a co-receptor, CD8 can bind to the alpha3 domain of the same MHC class I molecules as the TCR to facilitate TCR signaling. To evaluate the role of the MHC class I molecule alpha3 domain in the activation of CD8(+) CTL, we have produced a soluble 227 mutant of H-2D(d), with a point mutation in the alpha3 domain (Glu227 --> Lys). 227 mutant class I-peptide complexes were not able to effectively activate H-2D(d)-restricted CD8 T cells in vitro, as measured by IFN-gamma production by an epitope-specific CD8(+) CTL line. However, the 227 mutant class I-peptide complexes in the presence of another MHC class I molecule (H-2K(b)) (that cannot present the peptide) with a normal alpha3 domain can induce the activation of CD8(+) CTL. Therefore, in order to activate CD8(+) CTL, the alpha3 domain of MHC class I does not have to be located on the same molecule with the alpha1 and alpha2 domains of MHC class I. A low-avidity CD8(+) CTL line was significantly less sensitive to stimulation by the 227 mutant class I-peptide complexes in the presence of the H-2K(b) molecule. Thus, low-avidity CTL may not be able to take advantage of the interaction between CD8 and the alpha3 domain of non-presenting class I MHC molecules, perhaps because of a shorter dwell time for the TCR-MHC interaction.     

Characterization of in vivo expanded OspA-specific human T-cell clones

A panel of CD4 T-cell clones was isolated from synovial fluid by single cell flow cytometry from a patient with treatment-resistant Lyme arthritis using a DRB1*0401 major histocompatibility complex (MHC) class II tetramer covalently loaded with outer surface protein A (OspA) peptide164-175, an immunodominant epitope of Borrelia burgdorferi. Sequencing of the T-cell receptors of the OspA reactive clones showed significant skewing of the T-cell receptor repertoire. Of the 101 T-cell clones sequenced, 81 possessed TCR beta chains that were present in at least one other clone isolated. Complete sequencing of both alpha and beta chains of a subset of clones showed that at least two distinct T-cell clones were expanded in vivo. Binding studies using a panel of Ala-substituted peptide ligands were performed to determine potential MHC binding sites of the OspAp164-175 to DRB1*0401. In addition, T-cell clones were tested functionally for their reactivity to the wild-type peptide as well as to altered peptide ligands (APLs) and peptide libraries based on the OspA epitope in order to determine the TCR contact residues and the stringency in T cell recognition. We are among the first to define the characteristics of TCR usage of T cells isolated from an inflamed immune compartment in an individual with an autoimmune disease potentially triggered by a microbial antigen.     

Control of self-reactive cytotoxic T lymphocytes expressing γδ T cell receptors by natural killer inhibitory receptors

The majority of peripheral blood γδ T cells in human adults expresses T cell receptors (TCR) with identical V regions (Vγ9 and Vδ2). These Vγ9Vδ2 T cells recognize the major histocompatibility complex (MHC) class I-deficient B cell line Daudi and broadly distributed nonpeptidic antigens present in bacteria and parasites. Here we show that unlike αβ or Vγ9^? γδ T cells, the majority of Vγ9Vδ2T cells harbor natural killer inhibitory receptors (KIR) (mainly CD94/NKG2A heterodimers), which are known to deliver inhibitory signals upon interaction with MHC class I molecules. Within Vγ9δ2 T cells, KIR were mainly expressed by clones exhibiting a strong lytic activity against Daudi cells. In stark contrast, almost all Vγ9Vδ2 T cell clones devoid of killing activity were KIR^?, thus suggesting a coordinate acquisition of KIR and cytotoxic activity within Vγ9Vδ2 T cells. In functional terms, KIR inhibited lysis of MHC class I-positive tumor B cell lines by Vγ9Vδ2 cytotoxic T lymphocytes (CTL) and raised their threshold of activation by microbial antigens presented by MHC class I-positive cells. Furthermore, masking KIR or MHC class I molecules revealed a TCR-dependent recognition by Vγ9Vδ2 CTL of ligands expressed by activated T lymphocytes, including the effector cells themselves. Taken together, these results suggest a general implication of Vγ9Vδ2 T cells in immune response regulation and a central role of KIR in the control of self-reactive γδ CTL.     

Nef-mediated MHC class I down-regulation unmasks clonal differences in virus suppression by SIV-specific CD8 T cells independent of IFN-γ and CD107a responses

CD8⁺ T lymphocytes (CTL) play a role in controlling HIV/SIV infection. CTL antiviral activity is dependent on recognition of antigenic peptides associated with MHC class I molecules on infected target cells, and CTL activation can be impaired by Nef-mediated down-regulation of MHC class I molecules. We tested the ability of a series of rhesus macaque CD8⁺ T-cell clones specific for the SIV Gag CM9 peptide to suppress SIV infection of autologous CD4⁺ T cells. We used a set of SIV_mac239 viruses with either wild-type Nef or Nef mutations that impair MHC class I down-regulation. All CTL clones efficiently suppressed virus replication in cells infected with mutant viruses with altered Nef function, phenotypically MHC class I^high or MHC class I^intermediate. However, the ability of the clones to suppress virus replication was variably reduced in the presence of wild-type Nef (MHC class I^low) despite the observations that all CTL clones showed similar IFN-γ responses to titrated amounts of cognate peptide as well as to SIV-infected cells. In addition, the CTL clones showed variable CD107a (CTL degranulation marker) responses that did not correlate with their capacity to suppress virus replication. Thus, the clonal differences are not attributable to TCR avidity or typical effector responses, and point to a potential as yet unknown mechanism for CTL-mediated suppression of viral replication. These data emphasize that current assays for evaluating CTL responses in infected or vaccinated individuals do not fully capture the complex requirements for effective CTL-mediated control of virus replication.     

Control of TCR V alpha-mediated positive repertoire selection and alloreactivity by differential J alpha usage and CDR3 alpha composition

In rats expressing the f allele of the rat MHC (RT1f), CD8 T cells utilizing the V alpha 8.2 segment are 10-fold overselected during thymic development, resulting in V alpha 8.2 expression by 14% of mature CD8 T cells as compared to 1-2% in MHC congenic strains. In the alloreactive responses of CD8 T cells from RT1f-negative rats against RT1f, V alpha 8.2+ CD8 T cells are also preferentially expanded. Neither overselection nor alloreactivity of V alpha 8.2+ TCR require selective V beta pairing. However, RT1f alloreactive V alpha 8.2+ TCR preferentially use a related set of J alpha segments which contribute short homogeneous CDR3 alpha loops, with features suggesting peptide promiscuity, and little N additions. In contrast, only few overselected V alpha 8.2+ CD8 T cells showed an imprint of positive selection on J usage or CDR3 composition. The results demonstrate that a single V alpha segment can promote both MHC allele-specific positive selection and alloreactivity, and that the latter is more dependent on an additional contribution of CDR3 alpha, possibly by promoting reactivity with a diverse set of MHC-bound peptides or by providing additional MHC contacts.      

Distinct orientation of the alloreactive monoclonal CD8 T cell activation program by three different peptide/MHC complexes

We have characterized three different programs of activation for alloreactive CD8 T cells expressing the BM3.3 TCR, their elicitation depending on the characteristics of the stimulating peptide/MHC complex. The high-affinity interaction between the TCR and the K(b)-associated endogenous peptide pBM1 (INFDFNTI) induced a complete differentiation program into effector cells correlated with sustained ERK activation. The K(bm8) variant elicited a partial activation program with delayed T cell proliferation, poor CTL activity and undetectable ERK phosphorylation; this resulted from a low-avidity interaction of TCR BM3.3 with a newly identified endogenous peptide, pBM8 (SQYYYNSL). Interestingly, mismatched pBM1/K(bm8) complexes induced a split response in BM3.3 T cells, with total reconstitution of T cell proliferation but defective generation of CTL activity that was correlated with strong but shortened ERK phosphorylation. Crystal structures highlight the molecular basis for the higher stability of pBM8/K(bm8) compared to pBM1/K(bm8) complexes that exist in two conformers. This study illustrates the importance of the stability of both peptide/MHC and peptide/MHC-TCR interactions for induction of sustained signaling required to induce optimal CTL effector functions. Subtle allelic structural variations, amplified by peptide selection, may thus orient distinct outcomes of alloreactive TCR-based therapies.     

Growth-dependent variation of major histocompatibility complex (MHC) restriction and expression of Ly-2 and CD3/alpha/beta T cell receptor in cloned cytotoxic T cells

The specificity of major histocompatibility complex (MHC)-restricted antigen recognition by cytotoxic T cells (CTL) has been clearly correlated to the alpha/beta T cell receptor (TcR) complex on the T cell surface. Occasional changes in the specificity of in vitro cultivated CTL clones, therefore, have been suspected to result from alterations of the genes coding for the TcR alpha and/or beta chain. Here we demonstrate that pronounced variations in the stringency of MHC restriction, previously reported to occur during long-term culture of 2,4,6-trinitrophenyl (TNP)-specific CTL clones, may occur rapidly in a growth-dependent, reversible manner, i.e. without structural TcR variation. Several H-2b TNP-specific CTL clones were shown to possess strong cross-reactivity for H-2k TNP target cells when seeded at low cell numbers, but exhibit reduced or undetectable cross-reaction to H-2k TNP in high-density cultures. Another clone revealed "heteroclitic" properties with significantly stronger cytotoxic activity towards allogeneic (H-2k) than syngeneic (H-2b) TNP-modified target cells. In this case dilute cultures appeared as exclusively allo-MHC restricted, whereas dense cultures were allo/self cross-restricted. In all instances these phenomena were accompanied by cell density-dependent quantitative changes in the expression of Ly-2 and T cell antigen receptor. CTL from dilute cultures had at least 2-fold higher surface concentrations of Ly-2 and CD3 antigens than cells from dense cultures while other surface markers such as Thy-1 or LFA-1 were completely identical. No such effects were observed for CTL clones exhibiting cell density-independent specificity patterns. We conclude from these findings that (a) the stringency of MHC restriction specificity may be significantly affected by the amount of expressed TcR and/or Ly-2 molecules, (b) CTL possess mechanisms to regulate Ly-2 and TcR expression and, hence, their MHC-restricted antigen recognition, and (c) the ability to regulate Ly-2 and TcR expression may be altered during prolonged culture of a CTL clone.