Quantitative Analysis of T-Cell Receptor β Variable-Gene Usage in Cutaneous Late-Phase Reactions: Implications for T-Lymphocyte Recruitment in Cutaneous Inflammation

Abstract

To determine if functionally distinct T-lymphocyte (T cell) subsets accumulate in late-phase immunoglobulin E-mediated reactions (LPR), we quantitatively analyzed the immunophenotype and the T-cell receptor β variable-gene (Vβ) repertoire of T cells in cutaneous LPR. Peripheral blood and skin biopsies were obtained 6 or 24 h after sensitive subjects were challenged with intradermal injections of grass pollen allergen (Ag) and control (C) solution. The frequency of cells expressing CD3, CD4, CD8, CD45RO, and CD25/mm² was determined by immunohistochemistry in nine subjects. Vβ usage was assessed by reverse transcription-PCR in five of nine subjects. A significantly greater frequency of CD3⁺ and CD45RO⁺ (memory) T cells was detected in Ag sites than in C sites at 24 h after challenge but not at 6 h. The frequency of activated (CD25⁺) and helper (CD4⁺) T cells appeared to be increased in Ag sites as well, though not significantly. Vβ6 was the most commonly expressed Vβ detected in Ag sites, but it was also detected in accompanying C sites. Vβ2 was the most commonly expressed Vβ detected in C sites. Sequence analysis in one case revealed Vβ expression in a 6-h Ag site to be essentially polyclonal. Our findings suggest that memory T cells with Vβ expression similar to that in normal skin accumulate in developing cutaneous LPR. The limited usage of Vβ suggests a preferential recruitment or retention of reactive T cells from an endogenous subset of skin-homing T cells with its own skewed Vβ repertoire.

Late-phase reactions (LPR), occurring hours after human immediate immunoglobulin E-mediated responses in the skin and the respiratory tract, are postulated to play a pathogenic role in chronic allergic diseases (11). Cutaneous LPR generally occur only in sites of prior prominent mast cell (MC) activation. However, MC activation is likely insufficient by itself (18). MC mediators histamine, prostaglandin D₂ (18), tumor necrosis factor (7), interleukin 5 (IL-5), (a potent eosinophil activator) (9), and leukotriene C₄ induce skin inflammation but not a typical LPR. Thus, the exact link between MC activation and subsequent LPR requires further investigation.

Leukocytes accumulating in sites of LPR are generally thought to play a pathogenic role in this reaction and in chronic allergic diseases. Skin biopsy studies (26) show an initial neutrophil entry starting within 1 to 2 h, followed by eosinophils, basophils, and later, lymphocytes (5, 26). We found a correlation between the intensity of the gross skin LPR and the degree of local lymphocyte accumulation (1). Activated CD4⁺ T lymphocytes (T cells) have been demonstrated in cutaneous LPR 24 h after intradermal challenge (5). The stimuli for such accumulation and the role of the increasing number of T cells accumulating in the LPR site between 6 and 24 h after allergen (Ag) challenge are not known. Based upon the in situ hybridization findings of an increased frequency of T cells expressing mRNA for granulocyte-macrophage colony-stimulating factor, IL-3, IL-4, and IL-5 in such 24 h LPR (10), it has been postulated that T cells may secrete cytokines which activate locally accumulated granulocytes.

Evidence supporting the active participation of T cells which accumulate in sites of LPR is crucial to an understanding of their pathogenic role. Therefore, it is important to determine the factor(s) responsible for the accumulation of T cells and the specificity of the recruited T cells, particularly with regard to their interaction with the challenge antigen inducing the LPR. Studies of delayed hypersensitivity have indicated that the frequency of antigen-specific T cells within the lymphocytic infiltrate is small (<1 to 3%) (6, 8, 19, 21). Assessment of T-cell receptor (TCR) β variable-gene (Vβ) expression has been utilized to characterize T-cell infiltrates in inflammatory and neoplastic diseases (3, 16, 23). Based upon evidence of constrained usage of Vβ genes in T-cell responses to simple and complex antigens (15), analysis of Vβ usage would help determine whether there is an increased local frequency of T-cell clones reactive to the challenge antigen(s) at LPR sites. Therefore, in this study we sought to further characterize the function and possible pathogenic role of locally accumulating T cells found in developing and established cutaneous LPR. We quantitatively analyzed the immunophenotype (CD3, CD4, CD8, CD25, and CD45RO) and the Vβ repertoire of T cells accumulated in the skin after pollen Ag challenge.

MATERIALS AND METHODS

Subjects and sampling.

Nine adult subjects sensitive to grass or ragweed pollen extract were recruited for study after their informed consent was obtained. Subjects were challenged with intradermal injections (100 protein nitrogen units [PNU]/ml) of grass or ragweed pollen extract (Greer Labs, Lenoir, N.C.) (Ag) and diluent control solution (C). Replicate 3-mm-diameter punch biops of the skin were obtained from Ag and C sites 6 and/or 24 h after challenge while gross LPR were still present on the upper arm. In five of nine subjects, one additional 3-mm-diameter biopsy was obtained and snap frozen in liquid nitrogen for Vβ analysis by reverse transcription (RT)-PCR. Peripheral blood was obtained via venipuncture at the time of intradermal injections.

Immunohistochemistry.

Skin specimens were embedded in O.C.T. (Miles Inc., Elkhart, Ind.), snap frozen in liquid nitrogen, and stored at −70°C. Six-micrometer-thick frozen sections were obtained, fixed in acetone at 4°C, and then incubated with monoclonal antibodies (MAb) and the appropriate reagents of the APAAP (Dako, Carpinteria, Calif.) or avidin-biotin peroxidase techniques (ABC Vectastain kit; Vector Labs Inc., Burlingame, Calif.). The frequency of cells expressing CD3, CD4, CD8, CD25, and CD45RO/square millimeter of dermis was determined for all nine subjects. Comparisons were analyzed by the paired t test and Wilcoxon methods, depending whether the findings were normally distributed or not. Anti-CD3, anti-CD4, anti-CD8, anti-CD25 (IL-2 receptor), and anti-CD45RO MAb were from Becton Dickinson (San Jose, Calif.).

Isolation of RNA and cDNA synthesis.

Total RNA was isolated from skin and Ficoll gradient-isolated peripheral blood mononuclear cells (PBM) by standard techniques previously described (12). One to three micrograms of total RNA isolated from skin or blood (PBM) was reverse transcribed with oligo(dT) (Pharmacia Biotech, Piscataway, N.J.) with Superscript reverse transcriptase according to manufacturer recommendations (Promega, Madison, Wis.) in a total volume of 50 μl.

PCR.

Resultant first-strand cDNA was amplified according to the methods of Weidmann et al. (24). Two-microliter aliquots of the resultant first-strand cDNA were loaded into 22 separate PCRs, each with 100 ng of a primer pair specific for individual Vβs (Vβ1 through Vβ20) (4). Also added was 100 ng of a primer pair for the constant region of the TCR alpha gene (C_α) (24), which served as an internal standard. All primers were synthesized at the DNA Synthesis Facility at the University of Pennsylvania Cancer Center. Each PCR was run in a total volume of 50 μl and contained standard buffer (Perkin-Elmer Cetus, Norwalk, Conn.), 1.5 mM MgCl₂, 2.5 mM concentrations of each deoxynucleotide triphosphate with 10 μCi of [α-³²P]dCTP (Amersham, Arlington Heights, Ill.), primers, 1% gelatin (Sigma, St. Louis, Mo.), and 1.25 U of Taq polymerase (Perkin-Elmer Cetus). Cycle times were 30 s at 95°C, 1 min at 55°C, and 1 min at 72°C, with a total of 30 cycles. A single water blank (a PCR with all reagents except for template) was included in each panel of 22 Vβ reaction mixtures as a negative control. A single Vβ primer pair was randomly selected for each water blank.

Quantification of PCR product.

Amplification products were sized on a 3% acrylamide gel. Dried gels were scanned on a PhosphoImager (Molecular Dynamics, Sunnyvale, Calif.) after 40-min exposures, and comparative samples were exposed at the identical times. Each Vβ value was defined as the Vβ signal intensity divided by the C_α signal intensity amplified in the same tube (Vβ/C_α). Each Vβ value was then expressed as a percentage of the sum of the values for all the Vβ/C_α signal intensities.

Overexpression of a particular Vβ gene in a skin biopsy was defined as a value for that Vβ gene either (i) greater than 10% of the total Vβ values in the skin biopsy and at least twice that of PBM or (ii) greater than 30% of the total Vβ values in the skin biopsy (3).

DNA sequencing.

PCR products were subcloned into TA cloning vector (Invitrogen, San Diego, Calif.), and multiple individual subclones were sequenced by the dideoxynucleotide method (20).

RESULTS

CD3⁺ and CD45RO⁺ cells accumulate at 24 h Ag sites.

We found a modestly but not significantly greater number of total T cells (CD3⁺) in the Ag sites 6 h after intradermal challenge (when LPR were already grossly prominent) than in the C sites injected 6 h earlier with buffer diluent (55 ± 16 [mean ± standard error of the mean] versus 34 ± 7, P = 0.37) (Table 1). The phenotypic profiles of T-cell subsets (CD4⁺, CD8⁺, CD45RO⁺, and CD25⁺) at 6 h were also not significantly different at Ag and C challenge sites (Table 1). In contrast, there was a significantly greater frequency of CD3⁺ and CD45RO⁺ cells present in Ag challenge sites than in buffer C sites at 24 h (111 ± 19 versus 41 ± 9, P < 0.001; 67 ± 17 versus 42 ± 13, P < 0.05, respectively) (Table 1). Our finding of an increased frequency of CD45RO⁺ (putative primed or memory T cells) with only a relatively small proportion of T cells expressing the activation marker CD25 is in keeping with previous findings (5, 25).

TABLE 1.

Immunophenotypic profile of skin biopsies of LPR subjects

Immuno-phenotype	6 h biopsya			24 h biopsya
	Ag site	C site	P value	Ag site	C site	P value
CD3b	55 ± 16	34 ± 7	0.37	111 ± 19	41 ± 9	<0.001
CD4b	42 ± 8	31 ± 6	0.62	88 ± 20	41 ± 9	0.62
CD8b	11 ± 6	8 ± 3	0.273	9 ± 3	5 ± 2	0.15
CD25c	12 ± 8	7 ± 2	0.42	11 ± 4	7 ± 3	<0.08
CD45ROb	40 ± 4	35 ± 15	0.43	67 ± 17	42 ± 13	<0.05

^aMean number of cells/square millimeter ± standard error of the mean. 

^bAnalyzed by paired t tests. 

^cAnalyzed by the Wilcoxon method. 

Vβ6 is most commonly overexpressed in Ag sites, and Vβ2 is most commonly overexpressed in C sites.

We next sought to determine whether there was overexpression of particular Vβ genes in Ag and C challenge sites during the development of LPR. A total of 16 skin biopsies (8 Ag and 8 C sites) were obtained from five subjects and utilized for RT-PCR analysis. Through densitometric quantification of Vβ and Cα signal intensities, Vβ values were generated and overexpression was defined (as described in Materials and Methods). Table 2 summarizes the results. In three 24 h C site biopsies (subjects 3 to 5) no Vβ values were generated because none of the Vβ or Cα bands were amplified; therefore, these time points do not appear in Table 2. We attribute these findings to PCR detection sensitivity; i.e., the low level of TCR transcripts resulting from the paucity of lymphoid infiltrate within C site skin could not be amplified. In only one Ag site biopsy (subject 2, 24 h Ag), multiple Vβ genes (along with the C_α internal standard) did not amplify (for technical reasons), precluding generation of Vβ values and omission of this time point in Table 2. One to three Vβ genes were found to be overexpressed in the individual 6 h and/or 24 h Ag challenge sites in the five subjects. One to four Vβ genes were overexpressed in C sites of these subjects. Vβ6 was most frequently overrepresented in Ag site biops at 6 or 24 h Ag sites, seen in four of five subjects (1 to 4). In two subjects (1 and 3) in which both 6 and 24 h Ag biops were analyzed, Vβ6 was detected at both time points. Vβ6 was also overexpressed within the C site biops in both subjects. Vβ2 was most frequently overrepresented in C site biops (three of five) and was detected in both 6 and 24 h C sites in subject 2.

TABLE 2.

Summary of TCR Vβ overexpression^a in skin

Subject	Antigen		Control
	6 h	24 h	6 h	24 h
1	6, 18b	3, 5.2, 6	1	2, 6
2	1, 6, 18	NR	2, 16, 18	2, 7
3	6, 13.2	1, 6	5.1, 6, 11, 13.2	NR
4	—	6	—	NR
5	—	5.2	—	NR

^aTCR Vβs listed have values >10% and 2× peripheral blood lymphocyte or >30% of total skin expression. NR, no result (see text); —, no biopsy obtained. 

^bSequence analysis demonstrated that two of seven Vβ6 clones had identical in-frame sequences and that none of the eight Vβ18 clones were identical. 

In only one sample was sufficient cDNA available for additional PCR and sequencing. To determine the clonal nature of the Vβ6 and Vβ18 genes overrepresented in the 6 h Ag site of subject 1 (Table 2), cDNA was reamplified with Vβ6 and Cβ primers and Vβ18 and Cβ primers in separate PCRs without C_α primers or radioisotopes. After subcloning, sequence analysis revealed that only two of seven Vβ6 clones had identical in-frame sequences and that none of the eight Vβ18 clones were identical (data not shown). Thus, Vβ6 and Vβ18 overexpression appears to be essentially polyclonal, without evidence of dominant monoclonality.

DISCUSSION

Our RT-PCR analysis defined overexpression of Vβ genes in both Ag and C challenge sites at 6 and 24 h in all samples analyzed. Although Vβ6 was the most frequently expressed Vβ gene in Ag sites, its overexpression in accompanying C challenge sites within the same subject precludes the interpretation that its presence is related to antigen specificity. Furthermore, sequence analysis from one Ag site, demonstrating polyclonal and limited oligoclonal expansion of overexpressed Vβ subfamilies, does not support an antigen-specific clonal expansion of T cells. Although the number of specimens analyzed is small, the significantly increased number of memory (CD45RO⁺) T cells (CD3⁺) seen at 24 Ag sites (compared to matched C sites) most likely reflects nonspecific recruitment and local activation of T cells in the site.

The finding of Vβ6 in accompanying C sites as well as an increased frequency of Vβ2 suggests that T cells expressing these subfamilies may normally accumulate to a greater degree in the skin than in peripheral blood. Studies with primers, PCR conditions, and quantitative analysis similar to those utilized in our study have found an increased frequency of Vβ2 and/or Vβ6 in normal skin (14, 16), limited oligoclonal expansion of Vβ6 in lesions of leprosy (23), and varying degrees of clonal and oligoclonal Vβ2 and Vβ6 expansions in lesions of psoriasis (13, 14, 22). No definitive conclusions can be drawn regarding mechanisms underlying any presumed selective TCR Vβ usage in the above-mentioned studies.

Several factors may have influenced our ability to identify antigen-specific T cells in developing LPR, based on the findings of restricted Vβ usage. The pollen extracts approved for in vivo skin testing in humans are relatively crude, each containing a number of allergenic epitopes. Some subjects may be sensitized to more than one epitope, and a polyclonal T-cell response is thus induced. Limited amounts of RNA from 3-mm-diameter punch biops of the skin precluded complete sequence analysis in each subject. The experimental approach used in the current study and similar studies in skin (13, 14, 16, 22, 23) uses an arbitrary definition of Vβ overexpression based on the assumption that the Vβ repertoire in the skin normally mirrors the repertoire in the blood. Skin-homing T cells are defined by the surface expression of the cutaneous lymphocyte-associated antigen (CLA), a homing receptor that mediates cutaneous T-cell trafficking (2, 17). Defining Vβ overexpression by comparing the repertoire of skin-infiltrating T cells (a predominately CLA⁺ population) to that of peripheral blood (a predominately CLA⁻ population) may mask the detection of skin-homing subsets. At least one study in psoriasis has demonstrated that analysis of T-cell subsets within an inflammatory infiltrate may be more sensitive in detecting clonal T-cell subpopulations in the skin (3). Additional studies are needed to define the Vβ repertoire in CLA⁺ T cells and will aid future studies attempting to analyze Vβ usage in cutaneous inflammation.

In summary, our findings suggest that the memory T cells accumulating in developing cutaneous LPR display Vβ usage suggestive of nonspecific reactive cells recruited or retained from an endogenous T-cell population of the skin. The results of the current study and those of previous studies suggest that Vβ usage in cutaneous inflammation may be influenced by the inability to discriminate and compare functional T-cell subsets within the skin and blood. Further studies are required to fully define a restricted TCR Vβ repertoire in skin-homing T cells.