Immunology 1991 74 246-250
The genetic contribution to human T-cell receptor repertoire J. A.
LOVERIDGE, W. M. C. ROSENBERG, T. B. L. KIRKWOOD* & J.
BELL Molecular
ImmunologicGroup,
Institute of Molecular Medicine, John Radeliffe Hospital, Headington and *Laboratory of Mathematical Biology, Institute ftr Medical Research, The Ridgeway, Mill Hill, London
Accepted for publication
5
National
June 1991
SUMMARY Recent studies in mice have highlighted the importance of polymorphic genetic loci such as the major histocompatibility complex (MHC) or minor lymphocyte-stimulating antigen (Mls) in determining the nature of the peripheral T-cell receptor (TcR) population. ' I As our knowledge of the equivalent process in humans is incomplete, we have utilized a modification of the polymerase chain reaction (PCR)4 6 to determine the overall genetic contribution to the normal human TcR variable Vfl gene repertoire. These data demonstrate that the normal human T-cell population contains members of all the major TcR Vfl families and that there is considerable variation in the relative amounts of specific TcR Vf5 transcripts between individuals. We have established that the normal peripheral TcR V/# repertoire is more concordant in identical twins than in unrelated individuals. The relative importance of genetic factors in determining the peripheral TcR repertoire is emphasized by these results and suggests that, in humans, the genetic control of immune responsiveness is mediated in part by the peripheral TcR repertoire.
selection of T cells. One way to distinguish the contribution of genetic or environmental factors is to study the concordance of the TcR repertoire in adult identical twins and compare this to the repertoire in unrelated individuals.
INTRODUCTION
The TcR repertoire of mature circulating T cells is thought to be determined both by the genetic background of an individual and by clonal selection of T cells in response to environmental antigens. 237 Although genetic effects may arise through polymorphism of germline sequences,, it is likely that the selection of T cells during ontogeny and thus the character of the peripheral population of T cells is a programmed event dependent on genetically encoded molecules. In mice the T-cell receptor (TcR) has been shown to be significantly perturbed by both genetically encoded (I-E, Mls)'2,3'9 and environmental antigens (superantigens).4"'("" In man, little is known about the relative influence of environmental events or genetic factors on the repertoire of circulating T cells. Certain environmental superantigens, such as the toxic shock syndrome toxin I and the staphylococcal enterotoxins, are known to distort the TcR repertoire in man.'2 Exposure to such superantigens or to a range of other bacterial and viral pathogens may effect the TcR repertoire. Importantly, in man the long potential exposure time to these environmental events may mask any detectable genetic influence on the
MATERIALS AND METHODS
Tiw in identity The twins were all white Caucasians older than 25 years of age and resident in the U.K. Twin identity was confirmed using genomic Southern blot analysis with the variable number of tandem repeat (VNTR) probes, MS-1, MS-8, MS-31, MS-43 and MS-51 (obtained from A. Jeffries) (data not shown).
C(7-specific PCR Total RNA was isolated from peripheral blood lymphocytes (PBL) separated on Ficoll gradients, as described previously.5 Five micrograms of total RNA were then used for oligo-dTprimed cDNA synthesis5 and one-fifth of the cDNA diluted 10and 100-fold. Each of these dilutions was subsequently amplified in polymerase chain reactions (PCR) containing two TcR C/3-specific primers (5'-CCACACCCAAAAGGCCACACTGG-3' and 5'-GCTCTACCCCAGGCCTCGGCGC-3') at a final concentration of 0-2 JsM. The amplification was performed using 2 0 units of Taq polymerase (Perkin-Elmer, Emeryville, CA) and a Cetus/Perkin Elmer thermocycler under the following conditions; 95, I min; 60 , I min; 72 , 30 seconds, for a total of 25 cycles. Fifteen per cent of each PCR reaction was electrophoresed on a 2%, agarose gel, blotted to nitrocellulose (Hybond-C extra, Amersham, Amersham, Bucks, U.K.)
Abbreviations: Cf, constant region beta chain; Mis, minor lymphocyte-stimulating antigen; NOD, non-obese diabetic; PBL, peripheral blood lymphocytes; PCR, polymerase chain reaction; TcR, T-cell receptor; Va, variable region alpha chain; VfS. variable region beta chain; VNTR, variable number of tandem rcpeat. Correspondence: Dr J. 1. Bell, Molecular Immunology Group, Institute of Molecular Medicine, John Radcliffe Hospital, Headington OX3 9DU, U.K.
246
247
Genetic contribution to human TcR repertoire and probed with a radioactively labelled TcR C/3-specific oligonucleotide (5'-CGACCTCGGGTGGGAACAG-3').'3 The level of C/I-specific amplification was measured by densitometry on the autoradiograph utilizing a Chromoscan 3 (Joyce Loebl). By comparing the intensity of the C/ amplification product at all three different starting cDNA concentrations, the relative amounts of total TcR # template were determined to permit normalization of cDNA concentration before V/I amplification.
V/P oligo-specific PCR Initially, the concentration of total C/I cDNA was adjusted such that the amount of C/I template in each V/I-specific PCR reaction was identical (determined as described above). Parallel PCR reactions were then performed utilizing a TcR C/I-specific oligonucleotide (5'-CGCGAATTCAGATCTCTGCTTCTGATG-3') and a variable TcR VP-specific oligonucleotide (Table 1) at a final concentration of 0-2 pm and under identical conditions to those described for the TcR C/I PCR reactions. Again, 15% of each PCR reaction was electrophoresed on a 2% agarose gel, blotted to nitrocellulose and probed with a radioactively labelled TcR C/I-specific oligonucleotide. To provide a positive control for each VP family-specific PCR reaction, representative clones for each of the V/ families were obtained either by PCR on human cDNA made from PBL mRNA, or by screening a lambda gtlO library made from PBL mRNA. These were used in each experiment. A negative control, containing no template DNA, was also performed with each V/I-specific PCR reaction to control against contamination.
Data analysis The level of TcR V/I-specific amplification was measured by densitometry, as described above, and the difference between the concentration of individual V/ families in pairs of twins and in pairs of individuals calculated as a measure of discordance. The sum of these discordances was then subjected to a two-way analysis of variance to determine if the TcR repertoires of identical twins were more similar than those found in random individuals.
5'
Table 1. TcR V/I-specific oligonucleotides
V/Il: V/i2: Vfl3: V/14:
5'-CAACAGTTCCCTGACTTGCAC-3' 5'-TCAACCATGCAAGCCTGACCT-3' 5'-TCTAGAGAGAAGAAGGAGCGC-3' 5'-ACATATGAGAGTGGATTTGTCATT-3' Vfl5. 1: 5'-CTTCAGTGAGACACAGAGAAAC-3' V#5.2: 5'-CCTAACTATAGCTCTGAGCTG-3' Vfl6: 5'-GGCCTGAGGGATCCGTCTC-3' V/17: 5'-TGAATGCCCCAACAGCTCTC-3' V/18: 5'-ATTTACTTTAACAACAACGTTCCG-3' Vfl9: 5'-AATCTCCAGACAAAGCTCAC-3' V/hO: 5'-TCCAAAAACTCATCCTGTACCTT-3' V/ I 1: 5'-ACCAGTCTCCAGAATAAGGACG-3' V/l12: 5'-TGACAAAGGAGAAGTCTCAGAT-3' Vl13.1: 5'-GACCAAGGAGAAGTCCCCAAT-3' Vl13.2: 5'-TGGGTGAGGGTACAACTGCC-3' V/I14: 5'-CTCTCGAAAAGAGAAGAGGAAT-3' V/l 5: 5'-TCTCTCGACAGGCACAGGCT-3' V/I 16: 5'-AGAGTCTAAACAGGATGAGTCC-3' V/l 7: 5'-TCACAGATAGTAAATGACTTTCAG-3' V/I18: 5'-GAGTCAGGAATGCCAAAGGAA-3' V/il 9: 5'-CCCCAAGAACGCACCCTGC-3' V/120: 5'-TGAGGTGCCCCAGAATCTC-3' The sequences shown represent TcR V/I-specific sequences used in conjunction with the C/I primer (5'CGCGAATTCAGATCTCTGCTTCTGATG-3') to generate V/h family-specific PCR products.
RESULTS TcR fi chain cDNA PCR We have utilized PCR and a series of TcR V/I-specific primers to compare the relative amounts of V/I region family transcripts between different individuals.4 Because such a strategy might be liable to a variety of biases, special care has been taken to avoid such obstacles. Different sets of oligonucleotides amplify DNA sequences with varying efficiencies, therefore no attempt has been made to quantitatively compare the different TcR V region frequencies. Instead, the individual TcR frequencies were determined in pairs of individuals, either twins or random individuals. The concentration of TcR / chain cDNA was determined in all cDNA synthesized using a pair of TcR C/I-specific primers (Fig. 1). The total amount of C/ template was then equalized between the two samples to be amplified with the set of V/Ispecific primers to ensure that variation between individuals was not due to unequal starting quantities of TcR cDNA, or inhibitors of cDNA synthesis or amplification.
TcR Vfi PCR PCR was performed using a set of V/I-specific primers, each in conjunction with a single C/I-specific primer, to determine the frequency of each V/ family transcript in five sets of twins and two pairs of unrelated individuals (Figs 2 and 3). All twenty V/I
_ 9
1 J
I
_ 0 0
0
-
9
0
6
6o J
I
J
-
M
N
M
0 0
N
M
N
-
Figure 1. Autoradiograph of a Southern blot of amplified TcR C/I region illustrating the total amount of TcR # chain eDNA in each of the total cDNA syntheses from the individuals to be compared. I and J represent one set of identical twins compared to each other and M and N a second set of twins. Also shown for each set of TcR C/3 PCR reactions is a notemplate negative control (-). 1,5 Oligo-dT-primed cDNA from each individual was diluted 1 10, 1/100 and 1/1000 and amplified in an identical PCR reaction to that described in the Materials and Methods. utilizing two TcR C/I primers (5'-CCACACCCAAAAGGCCACACTGG-3' and 5'-GCTCTACCCCAGGCCTCGGCGC-3') at a final concentration of 0-2 pM. 15%, of each PCR reaction was electrophoresed on a 2%, agarose gel. blotted to nitrocellulose and probed with a radioactively labelled TcR C/I-specific oligonucleotide.
248
J. A. Loveridge et al.
Figure 2. Autoradiograph of a Southern blot of amplified TcR V/I regions illustrating the TcR V/I usage in the identical twins I and J. Shown for each TcR V/I PCR reaction are a positive control (+) consisting of the PCR amplification product of I ng of the appropriate full length TcR # chain cloned into the plasmid pBJ and a negative control (-).
Figure 3. Autoradiograph of a Southern blot of amplified TcR V/ regions illustrating the TcR V/ usage in the unrelated pair of individuals M and N. Also shown for each TcR V/I PCR reaction are a positive (+) control containing the PCR product of amplification from 1 ng of the appropriate full-length TcR chain in the plasmid pBJ and a notemplate negative control (-).
families were found in at least one individual. In addition, those V/ regions that were not detected after 25 cycles of PCR could be amplified by additional cycles of PCR or identified by anchor-PCR (data not shown).6
Table 2. Discordance values (variances) for paired densitometer readings
Measurement and analysis of TcR Vfi discordance To quantify these differences in intensity, each V/I-specific PCR product was measured by densitometry and each pair of readings transformed and tabulated, as shown in Table 2. Each row of figures in this table is a measure of the discordance between the amount of each TcR V/ family in the individuals being compared. To determine if the level of discordance of the TcR repertoire is higher in random individuals than in twins the data in Table 1 were subjected to a two-way analysis of variance yielding an F value of 7-78 and a P value < 0-001. This result indicated that the extent of discordance between densitometer readings did not vary significantly from band to band, but varied highly significantly among the seven pairs of individuals. On average, the discordances were markedly greater for the two
Source
Df
Subject pairs Bands Residual
6 1 382 0 2303 7 78 21 0 7076 0 0337 1-14 126 3 7267 0-0296
Total
153 5 8163
SS
MS
F
P
The genetic contribution to human T-cell receptor repertoire J. A.
LOVERIDGE, W. M. C. ROSENBERG, T. B. L. KIRKWOOD* & J.
BELL Molecular
ImmunologicGroup,
Institute of Molecular Medicine, John Radeliffe Hospital, Headington and *Laboratory of Mathematical Biology, Institute ftr Medical Research, The Ridgeway, Mill Hill, London
Accepted for publication
5
National
June 1991
SUMMARY Recent studies in mice have highlighted the importance of polymorphic genetic loci such as the major histocompatibility complex (MHC) or minor lymphocyte-stimulating antigen (Mls) in determining the nature of the peripheral T-cell receptor (TcR) population. ' I As our knowledge of the equivalent process in humans is incomplete, we have utilized a modification of the polymerase chain reaction (PCR)4 6 to determine the overall genetic contribution to the normal human TcR variable Vfl gene repertoire. These data demonstrate that the normal human T-cell population contains members of all the major TcR Vfl families and that there is considerable variation in the relative amounts of specific TcR Vf5 transcripts between individuals. We have established that the normal peripheral TcR V/# repertoire is more concordant in identical twins than in unrelated individuals. The relative importance of genetic factors in determining the peripheral TcR repertoire is emphasized by these results and suggests that, in humans, the genetic control of immune responsiveness is mediated in part by the peripheral TcR repertoire.
selection of T cells. One way to distinguish the contribution of genetic or environmental factors is to study the concordance of the TcR repertoire in adult identical twins and compare this to the repertoire in unrelated individuals.
INTRODUCTION
The TcR repertoire of mature circulating T cells is thought to be determined both by the genetic background of an individual and by clonal selection of T cells in response to environmental antigens. 237 Although genetic effects may arise through polymorphism of germline sequences,, it is likely that the selection of T cells during ontogeny and thus the character of the peripheral population of T cells is a programmed event dependent on genetically encoded molecules. In mice the T-cell receptor (TcR) has been shown to be significantly perturbed by both genetically encoded (I-E, Mls)'2,3'9 and environmental antigens (superantigens).4"'("" In man, little is known about the relative influence of environmental events or genetic factors on the repertoire of circulating T cells. Certain environmental superantigens, such as the toxic shock syndrome toxin I and the staphylococcal enterotoxins, are known to distort the TcR repertoire in man.'2 Exposure to such superantigens or to a range of other bacterial and viral pathogens may effect the TcR repertoire. Importantly, in man the long potential exposure time to these environmental events may mask any detectable genetic influence on the
MATERIALS AND METHODS
Tiw in identity The twins were all white Caucasians older than 25 years of age and resident in the U.K. Twin identity was confirmed using genomic Southern blot analysis with the variable number of tandem repeat (VNTR) probes, MS-1, MS-8, MS-31, MS-43 and MS-51 (obtained from A. Jeffries) (data not shown).
C(7-specific PCR Total RNA was isolated from peripheral blood lymphocytes (PBL) separated on Ficoll gradients, as described previously.5 Five micrograms of total RNA were then used for oligo-dTprimed cDNA synthesis5 and one-fifth of the cDNA diluted 10and 100-fold. Each of these dilutions was subsequently amplified in polymerase chain reactions (PCR) containing two TcR C/3-specific primers (5'-CCACACCCAAAAGGCCACACTGG-3' and 5'-GCTCTACCCCAGGCCTCGGCGC-3') at a final concentration of 0-2 JsM. The amplification was performed using 2 0 units of Taq polymerase (Perkin-Elmer, Emeryville, CA) and a Cetus/Perkin Elmer thermocycler under the following conditions; 95, I min; 60 , I min; 72 , 30 seconds, for a total of 25 cycles. Fifteen per cent of each PCR reaction was electrophoresed on a 2%, agarose gel, blotted to nitrocellulose (Hybond-C extra, Amersham, Amersham, Bucks, U.K.)
Abbreviations: Cf, constant region beta chain; Mis, minor lymphocyte-stimulating antigen; NOD, non-obese diabetic; PBL, peripheral blood lymphocytes; PCR, polymerase chain reaction; TcR, T-cell receptor; Va, variable region alpha chain; VfS. variable region beta chain; VNTR, variable number of tandem rcpeat. Correspondence: Dr J. 1. Bell, Molecular Immunology Group, Institute of Molecular Medicine, John Radcliffe Hospital, Headington OX3 9DU, U.K.
246
247
Genetic contribution to human TcR repertoire and probed with a radioactively labelled TcR C/3-specific oligonucleotide (5'-CGACCTCGGGTGGGAACAG-3').'3 The level of C/I-specific amplification was measured by densitometry on the autoradiograph utilizing a Chromoscan 3 (Joyce Loebl). By comparing the intensity of the C/ amplification product at all three different starting cDNA concentrations, the relative amounts of total TcR # template were determined to permit normalization of cDNA concentration before V/I amplification.
V/P oligo-specific PCR Initially, the concentration of total C/I cDNA was adjusted such that the amount of C/I template in each V/I-specific PCR reaction was identical (determined as described above). Parallel PCR reactions were then performed utilizing a TcR C/I-specific oligonucleotide (5'-CGCGAATTCAGATCTCTGCTTCTGATG-3') and a variable TcR VP-specific oligonucleotide (Table 1) at a final concentration of 0-2 pm and under identical conditions to those described for the TcR C/I PCR reactions. Again, 15% of each PCR reaction was electrophoresed on a 2% agarose gel, blotted to nitrocellulose and probed with a radioactively labelled TcR C/I-specific oligonucleotide. To provide a positive control for each VP family-specific PCR reaction, representative clones for each of the V/ families were obtained either by PCR on human cDNA made from PBL mRNA, or by screening a lambda gtlO library made from PBL mRNA. These were used in each experiment. A negative control, containing no template DNA, was also performed with each V/I-specific PCR reaction to control against contamination.
Data analysis The level of TcR V/I-specific amplification was measured by densitometry, as described above, and the difference between the concentration of individual V/ families in pairs of twins and in pairs of individuals calculated as a measure of discordance. The sum of these discordances was then subjected to a two-way analysis of variance to determine if the TcR repertoires of identical twins were more similar than those found in random individuals.
5'
Table 1. TcR V/I-specific oligonucleotides
V/Il: V/i2: Vfl3: V/14:
5'-CAACAGTTCCCTGACTTGCAC-3' 5'-TCAACCATGCAAGCCTGACCT-3' 5'-TCTAGAGAGAAGAAGGAGCGC-3' 5'-ACATATGAGAGTGGATTTGTCATT-3' Vfl5. 1: 5'-CTTCAGTGAGACACAGAGAAAC-3' V#5.2: 5'-CCTAACTATAGCTCTGAGCTG-3' Vfl6: 5'-GGCCTGAGGGATCCGTCTC-3' V/17: 5'-TGAATGCCCCAACAGCTCTC-3' V/18: 5'-ATTTACTTTAACAACAACGTTCCG-3' Vfl9: 5'-AATCTCCAGACAAAGCTCAC-3' V/hO: 5'-TCCAAAAACTCATCCTGTACCTT-3' V/ I 1: 5'-ACCAGTCTCCAGAATAAGGACG-3' V/l12: 5'-TGACAAAGGAGAAGTCTCAGAT-3' Vl13.1: 5'-GACCAAGGAGAAGTCCCCAAT-3' Vl13.2: 5'-TGGGTGAGGGTACAACTGCC-3' V/I14: 5'-CTCTCGAAAAGAGAAGAGGAAT-3' V/l 5: 5'-TCTCTCGACAGGCACAGGCT-3' V/I 16: 5'-AGAGTCTAAACAGGATGAGTCC-3' V/l 7: 5'-TCACAGATAGTAAATGACTTTCAG-3' V/I18: 5'-GAGTCAGGAATGCCAAAGGAA-3' V/il 9: 5'-CCCCAAGAACGCACCCTGC-3' V/120: 5'-TGAGGTGCCCCAGAATCTC-3' The sequences shown represent TcR V/I-specific sequences used in conjunction with the C/I primer (5'CGCGAATTCAGATCTCTGCTTCTGATG-3') to generate V/h family-specific PCR products.
RESULTS TcR fi chain cDNA PCR We have utilized PCR and a series of TcR V/I-specific primers to compare the relative amounts of V/I region family transcripts between different individuals.4 Because such a strategy might be liable to a variety of biases, special care has been taken to avoid such obstacles. Different sets of oligonucleotides amplify DNA sequences with varying efficiencies, therefore no attempt has been made to quantitatively compare the different TcR V region frequencies. Instead, the individual TcR frequencies were determined in pairs of individuals, either twins or random individuals. The concentration of TcR / chain cDNA was determined in all cDNA synthesized using a pair of TcR C/I-specific primers (Fig. 1). The total amount of C/ template was then equalized between the two samples to be amplified with the set of V/Ispecific primers to ensure that variation between individuals was not due to unequal starting quantities of TcR cDNA, or inhibitors of cDNA synthesis or amplification.
TcR Vfi PCR PCR was performed using a set of V/I-specific primers, each in conjunction with a single C/I-specific primer, to determine the frequency of each V/ family transcript in five sets of twins and two pairs of unrelated individuals (Figs 2 and 3). All twenty V/I
_ 9
1 J
I
_ 0 0
0
-
9
0
6
6o J
I
J
-
M
N
M
0 0
N
M
N
-
Figure 1. Autoradiograph of a Southern blot of amplified TcR C/I region illustrating the total amount of TcR # chain eDNA in each of the total cDNA syntheses from the individuals to be compared. I and J represent one set of identical twins compared to each other and M and N a second set of twins. Also shown for each set of TcR C/3 PCR reactions is a notemplate negative control (-). 1,5 Oligo-dT-primed cDNA from each individual was diluted 1 10, 1/100 and 1/1000 and amplified in an identical PCR reaction to that described in the Materials and Methods. utilizing two TcR C/I primers (5'-CCACACCCAAAAGGCCACACTGG-3' and 5'-GCTCTACCCCAGGCCTCGGCGC-3') at a final concentration of 0-2 pM. 15%, of each PCR reaction was electrophoresed on a 2%, agarose gel. blotted to nitrocellulose and probed with a radioactively labelled TcR C/I-specific oligonucleotide.
248
J. A. Loveridge et al.
Figure 2. Autoradiograph of a Southern blot of amplified TcR V/I regions illustrating the TcR V/I usage in the identical twins I and J. Shown for each TcR V/I PCR reaction are a positive control (+) consisting of the PCR amplification product of I ng of the appropriate full length TcR # chain cloned into the plasmid pBJ and a negative control (-).
Figure 3. Autoradiograph of a Southern blot of amplified TcR V/ regions illustrating the TcR V/ usage in the unrelated pair of individuals M and N. Also shown for each TcR V/I PCR reaction are a positive (+) control containing the PCR product of amplification from 1 ng of the appropriate full-length TcR chain in the plasmid pBJ and a notemplate negative control (-).
families were found in at least one individual. In addition, those V/ regions that were not detected after 25 cycles of PCR could be amplified by additional cycles of PCR or identified by anchor-PCR (data not shown).6
Table 2. Discordance values (variances) for paired densitometer readings
Measurement and analysis of TcR Vfi discordance To quantify these differences in intensity, each V/I-specific PCR product was measured by densitometry and each pair of readings transformed and tabulated, as shown in Table 2. Each row of figures in this table is a measure of the discordance between the amount of each TcR V/ family in the individuals being compared. To determine if the level of discordance of the TcR repertoire is higher in random individuals than in twins the data in Table 1 were subjected to a two-way analysis of variance yielding an F value of 7-78 and a P value < 0-001. This result indicated that the extent of discordance between densitometer readings did not vary significantly from band to band, but varied highly significantly among the seven pairs of individuals. On average, the discordances were markedly greater for the two
Source
Df
Subject pairs Bands Residual
6 1 382 0 2303 7 78 21 0 7076 0 0337 1-14 126 3 7267 0-0296
Total
153 5 8163
SS
MS
F
P
