Immunoprecipitation (IP) was performed using K562 cell lysate and associated RNA was extracted using phenol/chloroform/isoamyl alcohol (25:24:1) as described 1315. RNA pellets were resuspended in 30 µl RNA-grade water. cDNA was obtained from each RNA sample (10 µl) by reverse transcription (RT) using RT Master Mix (High Capacity cDNA RT kit, Applied Biosystems Inc., ABI, Foster City, CA, USA). The thermal cycler for the RT setting was: 10 minutes at 25°C, 120 minutes at 37°C, 5 seconds at 85°C. Quantitative PCR (qPCR) was performed using the TaqMan Fast Universal PCR Master Mix (ABI). For Th RNA (7-2 RNA, RMRP) detection, 'Hs03298751_s1' primer by ABI was used, while the primer for U3 RNA (SNORD3A, Integrated DNA Technologies, Coralville, IA, USA) has the following sequence:

Probe

Primer 1

Primer 2

In some experiments, La-depleted cell extract was also used to examine the effects of La depletion in a limited number of samples (n = 24). An extract from 25 × 10⁶ K562 cells was absorbed with 0.5 ml of anti-La immunoglobulin G-cyanogen bromide (IgG-CNBr)-activated Sepharose 4B beads to deplete La 16.

The protocol of this study was approved by the Institutional Review Board (IRB). This study meets and is in compliance with all ethical standards in medicine and informed consent was obtained from all patients according to the Declaration of Helsinki. For the titration experiment, one each of anti-Th/To, -U3RNP and normal human serum (NHS) was tested, with serial eight-fold dilutions (from 1 to 1:4096) of cell lysate as substrate in the IP step, RNA extracted after IP, and its derived cDNA. To validate the new assay, 22 anti-Th/To, 12 anti-U3RNP-positive samples, 58 SSc sera negative for anti-Th/To and -U3RNP antibodies (20 anti-Topo I, 18 anti-RNAPIII, 15 ACA, 5 anti-PM-Scl), 12 anti-La/SSB, 3 anti-trimethylguanosine (TMG) 17, and 15 NHS, were analyzed. The specificity of these sera was confirmed by IP and urea-PAGE analysis as described 13. Our data are representative of two or more independent experiments. PCR was analyzed in duplicate and immunoprecipitation was also tested in duplicate in some experiments. The typical reproducibility was very high showing a difference within one Ct value. In the new assay described here, no housekeeping gene mRNA would be present constantly, which is different from the standard qPCR used to examine the levels of mRNA for the gene of interest in total RNA or total mRNA. Nevertheless, the level of 18S RNA was also examined as a potential internal control for nonspecific binding (data not shown), but the results showed variation in its levels so that it was not considered suitable as an internal control in our assay.

Data between groups were compared by Kruskal-Wallis and Dunn's multiple comparison tests using Prism 5.0 for Macintosh (GraphPad Software, Inc). The same program was applied to obtain the receiver operating characteristic (ROC) curves, comparing the anti-Th/To or -U3RNP positive group versus the negative group to determine the cut-off. Statistical significance was accepted at P <0.05.

The sensitivity and linearity of the assay was evaluated for the detection of the Th and U3 RNA components, using serial 1:8 dilutions (1-1:4096) of cell lysate, RNA purified after IP (both started from extracts of 10⁷ cells per sample, but data shown are based on the use of 10 out of 30 µl of RNA, equivalent to 3.3 × 10⁶ cells) and cDNA obtained after RT (the equivalent of cDNA made from RNA derived from 3.3 × 10⁶ cells) (Figure 1). Ct values for the titration using serially diluted cell lysate were 8.3 (extract from 10⁷ cells, equivalent to undiluted or '1' in Figure 1A) to 22.7 (extract from 2.4 × 10³ cells, corresponding to 1:4096 in Figure 1A) for Th RNA and 13.4 to 28.9 for U3 RNA (not shown), using the prototype sera. Linear dose response curves were noted for cell lysate, RNA, and cDNA dilutions, and the difference between positive samples and NHS was clear even at 1:4096 dilution, which corresponded to a cell extract from 2.4 × 10³ K562 cells (Figure 1A). Ct values changed by approximately three for every 1:8 dilution, reflecting the eight-fold difference and the high efficiency of the qRT-PCR as expected (Figure 1A). Similar titration of RNA and cDNA, shown in Figure 1B and 1D, demonstrated the same linear relationship in each step of the qRT-PCR. IP using the standard 8 µl of human serum versus 1 µl was also compared in the titration experiment, to examine the effects on the Ct values. The difference in Ct values was approximately three, further demonstrating the high sensitivity of the new method also with 1 µl of serum.

The sensitivity of detection of Th or U3 RNA by the qPCR method was compared with that of the standard urea-PAGE-silver staining method by testing the same amount of serially diluted Th RNA (Figure 1B versus C) or U3 RNA (not shown). When using serial dilutions of RNA, the qPCR assay clearly detected a signal at the highest dilution of 1:4096 (Figure 1B) whereas Th RNA was barely visible only up to 1:16 dilution by silver staining (Figure 1C). Thus, qPCR was at least 256-fold more sensitive than silver staining. For U3 RNA detection, qPCR was clearly positive at a dilution of 1:4096, similarly to 7-2 RNA, whereas U3 RNA was detected only up to a 1:8 dilution by silver staining (data not shown), indicating that the qPCR assay was at least 512-fold more sensitive. Based on these data, cell extracts from one million cells, which correspond to 1/10th the amount of cell lysate normally used for IP, were used for screening of sera for anti-Th/To, -U3RNP, and other antibodies.

The mean Ct value was 19.81 (standard deviation (SD) 1.45) for anti-U3RNP and 15.27 (SD 0.73) for anti-Th/To samples (Figure 2A, B). These values were clearly separated from those of the control groups, with a difference of approximately eight Ct values for anti-U3RNP and approximately 13 Ct values for anti-Th/To (P <0.05) (Figure 2A, B). The Ct values of anti-TMG sera were similar to those of anti-U3RNP positive sera (Figure 2A). This was expected since anti-TMG antibodies recognize the cap structure on U1-U5RNA, including U3RNA 17. Despite maintaining statistical significance, the Ct difference between anti-Th and -La samples was smaller than the Ct difference between anti-Th and other negative controls (Figure 2C). This could be explained based on the interaction and co-immunoprecipitation of Th RNA by La antigens, as shown previously 18. For this reason, anti-La sera were also tested in the IP-qPCR assay using La-depleted K562 cell lysate. Ct values increased by about two and became comparable to other negative groups when La-depleted cell lysate was used in place of whole cell lysate (data not shown). This indicated that the low Ct values of anti-La sera in the anti-Th qPCR assay were due to co-IP of the Th RNA by the La complex, and not by coexisting anti-Th antibodies. When testing anti-U3RNP samples, 5/12 anti-La positive samples had low Ct values, which fell below the cut-off (Figure 2A). This might be explained by the observation that La could bind and stabilize precursors of the U3 small nucleolar RNA to prevent degradation 19. Thus, U3RNA could be weakly detected in the anti-La immunoprecipitate.

Considering IP with silver staining as the gold standard for the detection of anti-Th and -U3RNP antibodies, the sensitivity by our new IP-qPCR method was 100% for both (Table 1). Assay specificity for anti-Th and -U3RNP in SSc patients was 100% and 98.9%, respectively (Table 1). For anti-U3RNP detection, one anti-Th/To serum was repeatedly positive for an unknown reason.