Variation in the relative copy number of the TLR7 gene in patients with systemic lupus erythematosus and healthy control subjects




To determine whether there is an increase in the number of TLR7 gene copies in patients diagnosed as having systemic lupus erythematosus (SLE) and whether gene amplification influences the autoantibody profiles in SLE patients, as has recently been reported in the BXSB/Yaa mouse model of lupus.


We used a modified real-time quantitative polymerase chain reaction protocol to calculate the relative TLR7 gene copy number according to the comparative 2math image method in 99 SLE patients and 91 healthy controls matched for sex and ethnicity. Autoantibody profiles were determined by standard methods.


The relative number of TLR7 gene copies in SLE patients and healthy controls varied; however, no significant concordance between the number of relative gene copies and the SLE phenotype was found. There was also no difference in variation by ethnic group. Comparison of the relative gene copy numbers according to the presence or absence of antinuclear antibodies (ANAs), the ANA staining patterns, and the presence or absence of anti-RNA–associated antigen antibody showed no statistically significant difference in the SLE patients.


We determined that although the relative gene copy number of TLR7 varied in both SLE patients and healthy controls, it was not significantly increased among our SLE patients as compared with our controls. We found no detectable trend for an association between the relative gene copy number and the autoantibody profile in SLE patients.

A recent article in Science (1) reported that a genomic segmental duplication, which included the murine Toll-like receptor 7 (Tlr7) gene, and the translocation of this segment to the Y-linked autoimmune accelerator (Yaa) locus were associated with autoreactive B cell responses to RNA-related antigens. Similar findings of a duplication of the Yaa locus have been independently reported (2). The Yaa locus has previously been shown to increase the severity of lupus-like disease in males of the BXSB mouse strain (3) and to change their autoantibody specificity (4), leading to the suggestion that increased expression of Tlr7 due to this increase in genomic DNA may affect the autoimmune phenotype of these mice.

Mouse models can provide important insights into human immune function and disease; however, the mechanisms require careful validation, since many known immunologic differences exist between the two species (5). A role of TLR7 in humans with systemic lupus erythematosus (SLE) is consistent with its ability to induce the release of interferon-α (IFNα), a cytokine that has been shown to be increased in the serum of patients with SLE (6). Since TLR7 is located in a syntenic region of the X chromosome in humans and mice and since an increased prevalence of SLE in women (7) suggests an X-linked genetic component, we sought to determine whether, similar to the findings in the Yaa mouse, there were increased gene copy numbers of TLR7 in humans with SLE.


The relative number of copies of the TLR7 gene in genomic DNA from 50 Caucasian and 49 African-American patients with SLE (55 women and 44 men) along with 91 sex- and ethnicity-matched healthy control subjects was determined using a modified real-time quantitative polymerase chain reaction (PCR) method similar to that previously described for the detection of copy number polymorphisms in an immunologically related gene family (8). Samples were collected at The University of Alabama at Birmingham. All subjects gave their informed consent, and the Institutional Review Board approved the study.

The SLE patients met the revised and updated criteria established by the American College of Rheumatology (9, 10). We collected laboratory clinical data from as early as 1998, when available, on 63 of our 99 SLE patients, which included results of tests for anti-DNA antibodies, anticardiolipin antibodies, antinuclear antibodies (ANAs; including reactivity pattern, such as speckled, homogeneous, or nucleolar), and RNA–associated antigen autoantibodies (including anti-SSA/Ro, anti-SSB/La, anti-Sm, and anti-RNP).

We performed PCR quantitations in quadruplicate for each of the 190 samples, using Assay-by-Design TaqMan primers and FAM-labeled minor groove binder probes (all from Applied Biosystems, Foster City, CA) for TLR7 (Xp22.3). Sequences were as follows: for the forward primer, 5′-CAGTATTGTGCTGTCTTTGAAATGTAAA-3′; for the reverse primer, 5′-TGGGCCCAATAGCATCAACT-3′; and for the probe, 5′-TTGATGTCTTCTCTTTCTC-3′. To ensure that only genomic DNA was detected, the primers were designed to span an intron–exon border, and each sample was treated with an RNA-degrading enzyme during the DNA extraction process. We controlled for differences in DNA concentration between samples by normalizing against hypoxanthine phosphoribosyltransferase 1 (HPRT1; Xq26.1) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; 12p13), both of which have been previously reported to be housekeeping genes (11).

Reactions were conducted on the ABI 7900HT system, and calculations were performed using the 2math image method (12, 13). Briefly, this method calculates the difference in cycle thresholds (the number of PCR cycles required to produce a set of fixed thresholds) between the gene of interest and a housekeeping gene (ΔCt). Subsequent calculations normalize the ΔCt of each sample to a calibrator that is assigned a relative expression value of 1.00 (ΔΔCt). Assuming that the amount of PCR product doubles with each successive PCR cycle, calculating the 2math image value will provide the relative amount of DNA initially available for amplification in each quantitative PCR run. Therefore, the 2math image method reveals differences in the relative gene copy numbers between the samples tested (12). A range for each expression value was calculated based on the standard deviation(s) of the ΔΔCt value, where 2math image is the lower limit and 2math image is the upper limit. All statistical measures were calculated with GraphPad Prism 4.03 software (GraphPad, San Diego, CA).


Although the relative number of copies of the TLR7 gene in patients with SLE and in healthy control subjects varied, no significant concordance between the relative number of gene copies and the SLE phenotype was found. There was also no difference in variation by ethnic group. Therefore, we conclude that there is not an increase in relative copy number of the TLR7 gene in our population of SLE patients as compared with our population of controls (Figure 1).

Figure 1.

Relative number of copies of the TLR7 gene in genomic DNA from male and female patients with systemic lupus erythematosus (SLE) and healthy controls. Relative gene copy number of TLR7 in genomic DNA refers to the ratio of the quantity of TLR7 produced in a quantitative polymerase chain reaction (PCR) compared with that of a housekeeping gene, either A,GAPDH or B,HPRT1, as calculated by the 2math image method. Each sample was compared with 1 sample that was selected as a calibrator, to allow comparison of the relative amounts between individuals; therefore, these values reflect the comparison of the relative amount and not the absolute copy number. Given that the amount of resulting PCR product corresponds to the amount of genomic template DNA in each subject, the data presented here show no difference in the quantitative level of TLR7 in genomic DNA from SLE patients and healthy controls. Normalization against a housekeeping gene corrects for any sample-to-sample variation in the amount of template genomic DNA present in the PCR product. Symbols represent individual subjects. Bars show the mean ± SD for each group.

Correlational analysis (P < 0.0001) of both housekeeping genes, which were independently used in this study to normalize the amount of DNA in each sample, demonstrated that either gene could be used separately to calculate relative expression levels of TLR7. We are also confident in the technical precision of this method because the average variation in cycle number for our PCR results, which were performed in quadruplicate for each of the 3 assays (TLR7, GAPDH, and HPRT1) on each of the 190 samples, was <1% (0.86%) in this measure of intraassay precision (13). Additionally, when this study was repeated on a selection of 22 samples and then a selection of 54 samples (27 with SLE and 27 controls) using all 3 assays on different days, there was a significant correlation between samples for the interassay data (P = 0.0016) (13).

Since the Tlr7 copy number in mice has been shown to influence the titer and pattern of antibodies against nuclear antigens (1), we additionally attempted to correlate the relative gene copy number of TLR7 with autoantibody presence, titer, and pattern. Stratifying the relative gene copy number results by both positive and negative ANA titers, by ANA patterns, and by anti–RNA–associated antigen antibody presence showed neither a statistically significant difference nor an unequal distribution. This finding led us to conclude that the relative gene copy number of TLR7 does not have a significant pattern or relationship to the autoantibody profile in patients with SLE (Figure 2).

Figure 2.

Autoantibody profile versus the relative TLR7 gene copy number normalized against HPRT1 in systemic lupus erythematosus patients, by antinuclear antibody (ANA) positivity (in a speckled or homogeneous fluorescence staining pattern), ANA negativity, RNA-associated antigen (RNA-AA) autoantibody (anti-SSA/Ro, anti-SSB/La, anti-Sm, and anti-RNP) positivity, and RNA-AA negativity, and in healthy controls. To eliminate any variation due to sex, only data for the relative copy number of the TLR7 gene normalized against HPRT1 were used in this analysis. The distribution of each pattern among the range of relative gene copy numbers presents no significant relationships, as determined by analysis of variance (P = 0.7514). Symbols represent individual subjects. Bars show the mean ± SD for each group.


While Pisitkun and colleagues (1) demonstrated an example of copy number variation in a gene that influenced a complex disease in mice, their specific finding of a genomic increase in Tlr7 in a murine model of lupus cannot be translated directly to humans with SLE. Our findings in SLE patients do not preclude a role for TLR7 genetic variants, since coding region and/or regulatory variants (single-nucleotide polymorphisms) in the TLR7 gene are largely unexplored. In addition, an increase in the amount of message RNA for TLR7 may influence SLE.

We detected variations in the relative copy number of the TLR7 gene among a portion of both the patient and controls samples tested in our study. Due to the presence of this variation in both groups, we could not statistically associate this observed structural variation directly with the SLE phenotype. However, given the complex genetic nature of this autoimmune disease and given that TLR7 has potential functional relevance to SLE (6), we cannot rule out the possibility that copy number variations in TLR7 may influence the genetic background susceptibility for SLE. Since additional genetic variants are associated with lupus-like disease in mice, along with a TLR7 gene copy number variation (1), such a variant in the presence of other genetic factors may have a contributive, additive effect on the human SLE phenotype in a subset of patients.

Another recent study has emphasized sex differences in TLR7 function that might influence the SLE phenotype. Peripheral blood lymphocytes (PBLs) from healthy women release more IFNα after TLR-7 stimulation as compared with PBLs from healthy men, an effect that was not seen after stimulation with TLR-9, another inducer of IFNα (14). The increase in IFNα among PBLs from women was not due to a defect in X chromosome inactivation (14). Our results validate the findings of that study, since a significant difference in the relative TLR7 gene copy number in genomic DNA normalized against GAPDH was observed between men and women (P = 0.0138). While TLR7 may influence the genetic background of SLE pathogenesis and contribute to the difference in disease prevalence between the sexes, such a contribution in humans cannot be directly attributed to an increase in gene copy number in a standardized quantity of genomic DNA as was seen in the Yaa mouse.


Dr. Kelley had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Kelley, Johnson, Kimberly, Edberg.

Acquisition of data. Kelley, Alarcón.

Analysis and interpretation of data. Kelley, Johnson, Alarcón, Kimberly, Edberg.

Manuscript preparation. Kelley, Edberg.

Statistical analysis. Kelley, Edberg.


We would like to thank Jan Dumanski for critical review of the manuscript, Amy Peterson for technical assistance, and S. Louis Bridges for use of the ABI 7900HT sequencer.