Main

IDDM is a multifactorial T cell-mediated autoimmune disease that predominantly occurs in childhood. Susceptibility to IDDM is determined by genetic predispositions such as the MHC and environmental factors but the genuine etiology finally needs to be elucidated(1). An attractive hypothesis of its pathogenesis is that inflammation causing irreversible destruction of the β cells in the pancreatic islets of Langerhans might be triggered by a superantigen agent leading to selective expansion of autoreactive T cells via stimulation of the Vβ7 receptor family, respectively(2). In general, superantigens are able to circumvent the normal mechanisms for T cell activation but may stimulate certain subsets of T lymphocytes with shared [Vβ] receptors(3). Recently, it was proposed that the env of a novel HERV designated as IDDMK1,222 encodes such a superantigen, which was considered to be a candidate autoimmune gene for IDDM(4). IDDMK1,222 was supposed to be a novel member of the family of HERV-K and closely related to HERV-K10(4).

To further evaluate the hypothesis of a possible role of IDDMK1,222 in children with IDDM, we studied genomic existence and gene expression of this novel retrovirus in blood leukocytes from 55 children with IDDM (aged 2-19 y, mean age 11.4 y) and 55 healthy controls (aged 1-28 y, mean age 10.8 y). We achieved a high degree of sensitivity and specificity by using real-time PCR (reverse transcription and DNA PCR), a new approach for quantitative PCR. In addition, the identity of the PCR products obtained was proven by direct sequencing and aligning the sequences to related human retroviral elements.

METHODS

Patients and material. We analyzed blood leukocytes and plasma from 55 children with IDDM (aged 2-19 y, mean age 11.4 y, SD 3.9 y) including 11 patients at time of disease onset and of 55 healthy controls (aged 1-28 y, mean age 10.8 y, SD 6.8 y) with a negative family history for IDDM. Despite a greater SD for the control series there were no significant differences between the groups concerning age distribution or mean age (p = 0.55). The sex distribution was equal for the patient group and the control group (25 females, 30 males each). The control group consisted in 47 children from the outpatient clinic and 8 young adults (19-28 y) from our laboratory. The outpatients were included if there was no acute infection or endocrine and metabolic disease. Their diagnoses were constipation, headache, concussion, syncope, abrasions, inguinal hernia, umbilical hernia, cryptorchism, familiar delayed onset of puberty, or mild psychomotor retardation, respectively. Duration of IDDM in patients was 0-144 mo (median 50 mo), HbA1C ranging from 5.7-14.0% (normal 4.0-6.0%), mean HbA1C was 8.4% in the whole study group and 7.8% without newly diagnosed children. The study had been reviewed by the local ethics committee and was performed according to the principles expressed in the Declaration of Helsinki after informed consent was obtained from parents and patients.

DNA and RNA approach. DNA and RNA were isolated using conventional kits (QIAamp Blood Kit and RNeasy Blood Mini Kit, Qiagen, Hilden, F.R.G.). The integrity of DNA and RNA (i.e. ribosomal RNA bands) was determined by photographing ethidium-stained gels under UV light. DNA was pretreated with RNase before PCR and RNA was pretreated with DNase before reverse transcription was performed as reported previously(5). For plasma RNA extraction, we also used a commercial kit (QIAamp viral RNA Kit, Qiagen, Hilden, F.R.G.) according to a protocol of Conrad et al.(4).

PCR. To quantify gradual differences in gene expression with a high degree of sensitivity and specificity we applied real-time PCR (ABI PRISM 7700 Sequence Detection System, Perkin Elmer, Foster City, CA). This method detects PCR product generation through a fluorogenic oligonucleotide (i.e. TaqMan probe) located within the amplicon defined by a gene-specific oligonucleotide primer set. Briefly, the TaqMan probe is labeled with two different fluorescent dyes, a reporter dye (FAM, 6-carboxy-fluorescein) and a quenching dye (TAMRA, 6-carboxy-tetramethyl-rhodamine). When the TaqMan probe is intact, the reporter fluorescence is absorbed by the quenching dye. During PCR, the hybridized TaqMan probe is cleaved by the exonuclease activity of Taq DNA polymerase (AmpliTaq Gold), resulting in an increase of fluorescence. Nonspecific amplification is not detected. The initial quantity of the target template is determined by analyzing the fluorescence intensity. The fewer cycles it takes to reach a detectable fluorescence signal, the greater is the initial concentration of the template(6). The sensitivity of our PCR assay was >102 copies, determined by endpoint dilution series of appropriate purified DNA from PCR products. Specificity of our PCR approach was confirmed by direct sequencing of PCR outputs. The results were normalized to the expression of GAPDH, an ubiquitously expressed "housekeeping gene."

Sequences of the primers for GAPDH were: 5′-GCCATCAATGACCCCTTCATT-3′ (sense), 5′-TTGACGGTGCCATGGAATTT-3′ (anti-sense), 5′-FAM- CCTCAACTACATGGTTTACATGTTCCAATATGATTCCAC-TAMRA-3′ (sense), respectively.

For real-time PCR of IDDMK1,222 genomic existence and gene expression we applied distinct sets of primers for its U3 region and env gene, respectively. The sequences of the U3 primers were as follows: 5′-CCTTTACCTTGTCTATGATGCAAACA-3′ (sense), 5′- TTTTTGAGTCCCCTTAGTATTTATTGATC-3′ (anti-sense), 5′-FAM-CTTTGTTCACGTGTTTGTCTGCTGACCCTCTC-TAMRA-3′ (sense).

The sequences of the env primers were as follows: 5′-ATCATTATCCTCCTATTTGCCTAGGG-3′ (sense), 5′-CTAAAACTTCTGTATTCTTTGATCCTTTA-3′ (anti-sense), respectively, and 5′-FAM-AATTGGTTGGTAGAAGTACCTACTGTCAGT-TAMRA-3′ (sense). PCR reactions contained at a final concentration: 300 nM forward and reverse primer, 200 nM TaqMan probe, 200 µM dATP, dCTP, dGTP, 400 µM dUTP, 0.025 U/µl AmpliTaq Gold, 0.01 U/µl uracil-N-glycosylase in a total volume of 25 µL. A total of 2.5 µL of reverse transcription mixture (500 ng RNA in 40 µL volume) was used for RNA-PCR, corresponding to 32 ng genomic DNA in DNA-PCR. The thermocycler parameters were 50° for 2 min (for carry-over prevention with uracil-N-glycosylase), 95° for 10 min (for inactivation of uracil-N-glycosylase and hot start PCR), followed by 40 cycles of 95°C for 15 s and 60° for 1 min. Reagents were achieved from Perkin Elmer/Applied Biosystems, Weiterstadt, Germany (TaqMan PCR Core Reagent Kit).

DNA sequencing. Sequencing of PCR products was performed as reported previously(5) using an automated capillary sequencer (ABI PRISM 310, Perkin Elmer). For DNA sequencing the following primers were used: 5′-GTCTGAAATATGGCCTCGT-3′ (for U3) and specific PCR primers, respectively.

Data analysis. Software analysis and alignment was done with PCGENE and Omiga 1.01 (both from Oxford Molecular, Oxford, UK), or Geniusnet (German Cancer Research Center, Heidelberg, F.R.G.). All data are shown as mean ± SEM. Parametric data were analyzed with a two-tailed t test. p < 0.05 was considered significant. Nonparametric correlation was tested using Spearman's correlation coefficient.

RESULTS

In contrast to previous data(4) we were able to detect expression of the U3 region of IDDMK1,222 in almost all blood leukocytes from diabetic and healthy children. Statistical analysis revealed no differences of U3/GAPDH ratio in DNA and RNA between the two groups (Fig. 1,A and C). However, DNA sequence analysis showed PCR products of the U3 region of IDDMK1,222 indistinguishable from related human endogenous retroviral elements (GenBank accession numbers AF012335, X72790, M12852, M12853, M12854, X82271, X82272). Using real-time PCR primers for the env region encoding a putative superantigen, the sequence found was shown to be IDDMK1,222 by DNA sequencing. The env/GAPDH ratios in human leukocyte DNA and RNA were similar for diabetic and healthy children (Fig. 1,B and D). In general, the expression of IDDMK1,222 env and U3 RNA in patients and controls was quite heterogeneous: some individuals presented inexplicable high RNA levels, independent of disease or duration of IDDM, whereas others had low RNA levels. U3 RNA/GAPDH ratio was 0.0024 ± 0.00065 (range: 0.000-0.021) for patients and 0.0023 ± 0.00072 (range: 0.000-0.018) for controls, respectively. Env RNA/GAPDH ratio was 0.0038 ± 0.00078 (range: 0.000-0.017) for patients and 0.0051 ± 0.00144 (range: 0.000-0.053) or controls.

Figure 1
figure 1

Real-time DNA-PCR and reverse transcription PCR for IDDMK1,222 genomic existence and gene expression. A, IDDMK1,222-U3/GAPDH ratio in DNA of human leukocytes. B, IDDMK1,222-ENV/GAPDH ratio in DNA of human leukocytes. C, IDDMK1,222-U3/GAPDH expression of RNA in human leukocytes. D, IDDMK1,222-ENV/GAPDH expression of RNA in human leukocytes. Bright bars at the left-hand side, diabetic children (n = 55), dark bars at the right-hand side, nondiabetic age-matched controls (n = 55). Values are given as the mean ± SEM of relative units (ratios). No significant differences were detected.

On DNA level, env gene presence correlated well with detection of the U3 region of IDDMK1,222 for controls and diabetic children (r = 0.6419, p < 0.0001 and r = 0.5609, p < 0.0001, respectively). This could not be confirmed on RNA levels for either group (r = 0.2178 and 0.0503, not significant). For our opinion his could be due to a different expression of the env and U3-R-poly (A) region of IDDMK1,222 or, more likely, due to distinct sensitivity and PCR effectivity of the applied primer sets.

IDDMK1,222 U3 and env DNA and RNA neither correlated with islet cell antibodies or insulin autoantibodies as parameters of autoimmune β cell destruction in our 11 recently diagnosed patients (data not shown) nor was any correlation with HbA1C found in the whole patient group (r = -0.02 and 0.09 for U3 and env DNA after normalization to the housekeeping gene GAPDH, r = 0.16 and 0.05 for U3 and env RNA after normalization to GAPDH, respectively). Additionally, in many plasma probes from both groups U3 and env gene expression was found. In two plasma samples obtained from diabetic children (one acute onset case and one long-standing patient) a high expression of U3 RNA and GAPDH RNA was found (relative ratio U3 RNA/GAPDH RNA 0.075 and 0.522), all other plasma specimens from patients were negative for U3 RNA but none for GAPDH RNA. In the control group four subjects had U3 RNA/GAPDH RNA ratios ranging from 0.0006 to 0.02. Env gene expression was found more often in the samples: 47% of the patients and 42% of the controls exhibited env RNA. The mean ratios env RNA/GAPDH RNA were 0.0017 (SEM 0.00064) for patients and 0.0014 (SEM 0.00040) for controls, respectively. Interestingly, expression of IDDMK1,222 env gene correlated quite well with the presence of GAPDH in plasma, indicating that IDDMK1,222 expression was most likely from lysed blood leukocytes (r = 0.63, p < 0.001). DNA contamination could be ruled out because all probes were pretreated with DNAse before reverse transcription was performed. Different expression pattern for U3-R-poly(A) RNA and env RNA in plasma samples may be due to a greater nuclease sensitivity of U3-R-poly(A) RNA.

Negative samples were shown to stay negative when analysis was repeated to rule out cross- over contamination. For four patients, the whole procedure including RNA extraction, reverse transcription, and real-time PCR was repeated 4 mo later with same results (data not shown).

DISCUSSION

Human endogenous retroviral elements are widely distributed all over the genome. HERV sequences are considered to encompass up to 1% of the human genome(7). Expression of human endogenous retrovirus RNA has been detected in several tissues and blood cells(810). HERVs are preferentially expressed in embryonic tissues, especially placenta(11). There is the theoretical possibility that endogenous retroviral elements may recombine with sequences of exogenous or other endogenous retroviruses, resulting in new infectious or oncogenic potential(12). Additionally, the defective human terato-carcinoma-derived virus is encoded by a human endogenous retroviral sequence and its expression is enhanced in patients with germ cell tumors(13).

In our study we examined the spontaneous expression of the env and U3 region of IDDMK1,222, a novel human endogenous retrovirus. Although there were differences among individuals, they could not be related to diabetes onset or duration. Our findings are in contradiction to former data(4), most likely due to methodical improvements such as pretreatment of samples with DNase or RNase, gradual detection of gene expression by real-time PCR and sequencing of DNA products, respectively. Our results, however, do not disprove the hypothesis that superantigens in general may be involved in the pathogenesis of insulin-dependent diabetes. Indeed, from our point of view the pathophysiological role of IDDMK1,222 seems to be rather low because it could be detected in diabetic and healthy children in comparable quantities. Nevertheless, the situation in type 1 diabetes may be more complex because HLA genes, especially HLA-DQ, are the main factors predisposing to IDDM. Recently it was suggested that endogenous retroviral long terminal repeat elements within the HLA-DQ region represent an additional susceptibility marker for IDDM(14). However, it could not be demonstrated convincingly that HERVs indeed play a pathophysiological role as causative agents in autoimmune diabetes mellitus. IDDM is the clinical end point of a cascade of immunologic events, beginning early in the maturation and activation of T cells(15), respectively. Therefore, further studies are necessary to elucidate autoimmune response leading to β cell destruction in humans. Recently, it has been shown by others(1618), using different molecular and immunological techniques that IDDM does not depend on specific expression of IDDMK1,222 and that IDDMK1,222 and HERV-K10-like viremias are common in healthy subjects, respectively. Additionally, we have performed a novel quantitative PCR approach in a large pediatric collective to disprove a fascinating hypothesis for the pathogenesis of IDDM. We could not find a convincing correlation between U3 and env region of IDDMK1,222 on RNA level, which may be explained by a different pattern of expression or, more likely, due to distinct sensitivity and PCR effectivity of the applied PCR primers, also we have to take into account that PCR of U3 region led to a mixture of highly similar sequences, a result confirmed by others(16).

However, because there is evidence for the functional behavior of IDDMK1,222 as a superantigen(19) this issue needs further investigation to identify the factors leading to activation of autoreactive T cells in type 1 diabetes.

Taken together we have set an example for the application of real-time PCR. Ruling out typical pitfalls such as DNA contamination this method opens a wide variety of applications especially in pediatric studies where gene expression in low quantities or small specimens, respectively, is particularly important.