SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Objective

To identify candidate genes that are regulated by human pregnancy and have the potential to modulate rheumatoid arthritis (RA) disease activity.

Methods

Peripheral blood mononuclear cells (PBMCs) from healthy pregnant volunteers were analyzed using Affymetrix GeneChips at 4 time points (during the first, second, and third trimesters and 6 weeks postpartum). Based on the GeneChip data, target genes were further analyzed via real-time quantitative polymerase chain reaction (qPCR) using PBMCs from healthy controls and RA patients. In order to determine the cellular source of the candidate gene messenger RNA (mRNA), monocytes and lymphocytes from healthy controls and RA patients were positively selected using magnetic beads, and their mRNA was analyzed by qPCR.

Results

One-way analysis of variance identified 1,286 mRNAs that were differentially expressed with regard to the 4 time points. The changes became more pronounced as pregnancy progressed, and they were reversed postpartum. A subsequent pathway analysis suggested a regulatory role of pregnancy on the adipocytokine pathway as well as on the peroxisome proliferator–activated receptor (PPAR) signaling pathway. Of 19 preselected candidate genes, AKT3, SOCS3, FADS2, STAT1, and CD36 proved to be differentially regulated by pregnancy. In samples from RA patients, the differences were concordant with those in healthy controls but more pronounced. Both T lymphocytes and monocytes contributed to the regulated expression of these genes.

Conclusion

Our findings indicate that normal human pregnancy leads to changes in the expression of several molecular pathways in PBMCs, which are reversed postpartum. Changes in RA patients, although concordant, exceed the levels observed in healthy controls. Genes of the adipocytokine and PPAR signaling pathways qualify as candidates for the modulation of RA disease activity during pregnancy.

Pregnancy is an exceptional immunologic state, as the mother's immune system has to avoid rejection of the semiallogeneic fetus but preserve immunocompetence to fight infection and eliminate transformed cells. A well-known observation in human pregnancies is the “spontaneous” clinical improvement of certain systemic autoimmune diseases, such as rheumatoid arthritis (RA) and multiple sclerosis (1, 2). This disease-remitting effect of pregnancy is reversed postpartum.

Several mechanisms that have been proposed to be operative in the tolerance of the mother toward the fetus are also involved in pregnancy-induced amelioration of RA (3, 4). Among these, antiinflammatory cytokines (5) as well as fetal–maternal HLA incompatibility (6) play important roles. Furthermore, it has been shown that an expansion of the Treg cell population is important for maternal tolerance toward the fetus and correlates with the improvement of RA during pregnancy (7, 8).

In a recent study using transcriptome data from peripheral blood mononuclear cells (PBMCs) obtained during and after pregnancy, we showed a decrease in the activity of T lymphocytes with preserved innate mechanisms during pregnancy (9). In RA patients, the postpartum disease relapse was associated with a persistence of the activity of innate immune cells and an increase in the expression of genes of T lymphocytes.

These data suggest that the positive effects of human pregnancy on autoimmune diseases are a result of the immunologic changes which originate at the fetal–maternal interface but become operative at the systemic level and finally reduce disease activity in the inflamed joints of RA patients. In order to further analyze which pathways and genes may be responsible for the described effects, we analyzed gene expression in the PBMCs of healthy women during and after pregnancy and compared these findings with findings in samples from RA patients. With the identification of individually regulated genes and pathways, novel strategies for the treatment of systemic autoimmune diseases such as RA might be envisioned.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Patients.

We prospectively analyzed blood samples from 2 cohorts of healthy women. Cohort 1 consisted of 32 individuals who were recruited in a cross-sectional manner during the first trimester (gestational weeks 11–14; n = 8), during the second trimester (gestational weeks 20–23; n = 8), and during the third trimester (gestational weeks 30–33; n = 8) of pregnancy as well as 6 weeks postpartum (n = 8). A prospective, longitudinal analysis of blood samples obtained from 8 healthy pregnant women (cohort 2) and from 8 pregnant RA patients (cohort 3) during the third trimester and postpartum was performed. RA disease activity was assessed with the 3-variable Disease Activity Score in 28 joints using the C-reactive protein level (DAS28-CRP) (10), where the 3 variables are swollen joint count, tender joint count, and CRP (11).

The study was approved by the ethics commission of the Canton of Bern, Switzerland. Written informed consent was obtained from all subjects.

Sample preparation.

Immediately after blood samples were obtained, PBMCs were isolated by Ficoll-Histopaque (Sigma-Aldrich) density-gradient centrifugation and washed 3 times at 4°C. For cohorts 2 and 3, the PBMC samples were divided into 3 aliquots. The first aliquot was used for immediate RNA extraction, and the second and third aliquots were used for magnetic bead cell separation to isolate T lymphocytes (CD3+) and monocytes (CD14+). Cell separation was performed according to the recommendations of the manufacturer (Miltenyi Biotec).

Total RNA was isolated using an RNeasy Mini kit according to the recommendations of the manufacturer (Qiagen). RNA quality was determined with an Agilent 2100 Bioanalyzer and an RNA 6000 Nano kit (Agilent).

GeneChip analysis.

Equal amounts of the RNA extracted from the samples in cohort 1 were pooled (to save costs) for each time point as follows. Pool A consisted of samples 1, 2, 7, and 8; pool B of samples 1–5, and pool C of samples 4–8. The RNA was transcribed and labeled according to the recommendations of the manufacturer (Affymetrix). It was then hybridized onto Affymetrix Human Gene 1.0 ST microarray chips. The hybridization cocktail (80 μl) with fragmented biotin-labeled target DNA at a final concentration of 25 ng/μl was transferred into the Affymetrix GeneChips and incubated at 45°C in a hybridization oven 640 (Affymetrix) for 17 hours at 60 revolutions per minute. Subsequently, arrays were washed, stained, and processed according to the recommendations of the manufacturer. Raw (.dat) image files of the microarrays were generated using Affymetrix GeneChip Command Console (version 0.0.0.676). GeneChip results were normalized using the robust multiarray average algorithm (12).

Real-time quantitative polymerase chain reaction (qPCR).

In order to confirm the results of GeneChip analysis, real-time qPCR experiments were performed. TaqMan assays were designed and ordered from Applied Biosystems for the following genes: AKT3 (Hs00178533_m1), CD28 (Hs00174796_m1), CD36 (Hs01567186_m1), CD3E (Hs00167894_m1), FADS2 (Hs00188654_m1), PPP3CB (Hs00236113_m1), SOCS3 (Hs01000485_g1), STAT1 (Hs01013990_m1), and GAPDH (Hs99999905_m1).

Two hundred fifty nanograms of total RNA was reverse transcribed in a reaction volume of 25 μl using reverse transcriptase and random hexamers as a primer (both from Promega). The PCRs were performed with an ABI 7500 LightCycler (Applied Biosystems) using TaqMan Fast Master Mix. Reaction mixtures contained 1 μl of complementary DNA aliquot and 2 μM relevant primers. The mean expression values of duplicate samples were normalized to the mean values obtained for the GAPDH housekeeping gene.

Statistical analysis.

Differences in gene expression were analyzed by one-way analysis of variance (ANOVA) using TM4 software version 4.4 (13), with an overall threshold P value of 0.01.

Co-occurrence discovery analysis (using GeneCodis) (14) and pathway mapping (using PathVisio version 1.1) (15) were performed to search for pathways with differentially regulated gene expression. P values for GeneCodis were obtained through hypergeometric analysis and corrected by the false discovery rate method. Clusters were created by hierarchical clustering (using GeNESiS 1.7.5) (16).

Further statistical analyses were performed using SPSS software version 19. Significant differences were calculated using the Wilcoxon signed rank test for longitudinal comparisons of paired samples. The Mann-Whitney U test was performed for unpaired data and groupwise comparisons. For all tests, P values less than 0.05 were considered significant. It should be noted that due to the many tests performed, true P values might lie above our nominal P values. Since we expected strong correlations of the test results, Bonferroni correction would be excessive, but P values that are only slightly lower than 0.05 should be interpreted with caution.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Results of GeneChip experiments.

First, we analyzed and compared the gene expression profiles of PBMCs from healthy women (cohort 1) during and after pregnancy. Using one-way ANOVA, we identified 1,286 genes that were differentially expressed among the 3 trimesters and at the postpartum time point (P < 0.01) (Figure 1). As expected, the expression profiles for the 3 trimesters were more closely related than were the pregnancy expression profiles and the postpartum expression profiles. Of note, the changes became more pronounced as pregnancy progressed, and they were reversed postpartum.

thumbnail image

Figure 1. Gene expression profiles during and after pregnancy in healthy women, shown by hierarchical clustering of 1,286 genes. Peripheral blood mononuclear cell samples were obtained from a total of 32 healthy women (cohort 1) in the first trimester (1T; n = 8), second trimester (n = 8), and third trimester (n = 8) of pregnancy and 6 weeks postpartum (pp; n = 8). GeneChip analysis was performed, and significant differences in gene expression levels (P < 0.01) were determined by one-way analysis of variance. Samples for each time point were grouped into 3 pools (A, B, and C).

Download figure to PowerPoint

Pathway analysis and selection of individual genes.

The co-occurrence discovery method using the 1,286 genes identified showed 17 significantly different Kyoto Encyclopedia of Genes and Genomes pathways (Table 1). As expected, and consistent with the results of our previous study (9), the T cell receptor signaling pathway showed a strong difference between the time points during pregnancy and the postpartum time point. This is further illustrated in Figure 2, which shows lower expression of messenger RNA (mRNA) for the T cell receptor CD3E during pregnancy than postpartum. Regarding pathways, the changes in activity found in the adipocytokine signaling pathway and the peroxisome proliferator–activated receptor (PPAR) signaling pathway were of particular interest. Recent data have shown that these pathways may play a role in the pathogenesis of and control of disease activity in RA (17, 18). Thus, we selected these 2 pathways for further analysis and used the T cell receptor signaling pathway as an internal control.

Table 1. Significantly enriched KEGG pathways among the 1,286 genes that were differentially regulated during pregnancy and postpartum in healthy women*
KEGG pathway no.Pathway nameNo. of differentially regulated genesTotal no. of genes in pathwayP
  • *

    KEGG = Kyoto Encyclopedia of Genes and Genomes; PPAR = peroxisome proliferator–activated receptor.

  • From co-occurrence discovery analysis using GeneCodis.

  • Selected for further analysis.

04660T cell receptor signaling pathway18931.719−5
05200Pathways in cancer353206.193−5
04920Adipocytokine signaling pathway13673.212−4
04640Hematopoietic cell lineage14814.376−4
05222Small-cell lung cancer14834.679−4
05216Thyroid cancer8296.835−4
05210Colorectal cancer13841.699−3
03320PPAR signaling pathway11683.491−3
04664Fcε receptor I signaling pathway11746.651−3
04010MAPK signaling pathway232622.194−2
04210Apoptosis11872.169−2
04650Natural killer cell–mediated cytotoxicity141292.211−2
05223Non–small-cell lung cancer8542.884−2
05014Amyotrophic lateral sclerosis8553.015−2
04620Toll-like receptor signaling pathway11963.503−2
04020Calcium signaling pathway161764.9231−2
05220Chronic myeloid leukemia9744.753−2
thumbnail image

Figure 2. Hierarchical clustering of genes chosen for their differential regulation during and after pregnancy in healthy women. Genes were selected based on their regulation patterns in the first trimester (1T), second trimester (2T), and third trimester (3T) and at 6 weeks postpartum (pp). Samples for each time point were grouped into 3 pools (A, B, and C). Targets chosen for further analyses are shown in boldface. P values were determined by one-way analysis of variance to analyze differences in gene expression between all 4 time points.

Download figure to PowerPoint

With the help of pathway mapping, and by comparing their regulation patterns during and after pregnancy as well as their statistical significance, 19 candidate genes were selected (Figure 2). Whereas mRNAs for SOCS3, STAT1, and CD36 were up-regulated, the other candidate genes were down-regulated in the third trimester compared to the postpartum time point.

Next, we used qPCR to confirm the results of the GeneChip analysis. As expected, there was not full concordance. Approximately 75% of the qPCR results corresponded with the GeneChip data (see Supplementary Figure 1, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131). Although the data were consistent with our expectations, we decided to verify these findings using mRNA samples from an independent cohort of healthy donors. The results obtained with the pooled mRNA samples were confirmed by qPCR of the individual samples (data not shown). Finally, the candidate genes with the best concordance and with the greatest regulatory potential were chosen for further analysis. These were SOCS3, STAT1, CD36, PPP3CB, AKT3, CD28, FADS2, and CD3E.

Longitudinal analysis and comparison of samples from healthy donors and RA patients.

We then compared the expression patterns of the selected genes in 2 additional cohorts, one consisting of healthy controls (cohort 2) and one consisting of patients with RA (cohort 3), that were longitudinally followed up during and after pregnancy (Figure 3). Of note, all RA patients experienced an amelioration of their disease during pregnancy and a postpartum reactivation as defined by changes in the DAS28-CRP (clinical data not shown).

thumbnail image

Figure 3. Comparison of gene expression levels in normal, healthy donors (ND) and patients with rheumatoid arthritis (RA) during and after pregnancy. The mRNA samples obtained from the peripheral blood mononuclear cells of healthy donors (n = 7) and RA patients (n = 8) in the third trimester (3T) of pregnancy and 6 weeks postpartum (pp) were analyzed by quantitative polymerase chain reaction. The relative expression values are shown as the fold change compared to GAPDH mRNA expression. Longitudinal changes within groups, differences between the groups at each time point, and the differences between the groups in the increase in the relative expression from the third trimester to the postpartum time point were analyzed. ∗ = P < 0.05 by Wilcoxon signed rank test for paired data; ++ = P < 0.01 by Mann-Whitney U test for independent samples; Δ = P < 0.05, relative expression from the third trimester to the postpartum time point in healthy donors versus RA patients, by Mann-Whitney U test for independent samples.

Download figure to PowerPoint

The differences in the expression of CD28, CD3E, STAT1, AKT3, and FADS2 during pregnancy versus after pregnancy in the healthy cohort were significant. The data were consistent with the cross-sectional data. Compared to healthy controls, RA patients showed significantly higher expression of CD28, PPP3CB, STAT1, AKT3, FADS2, and CD36 (P < 0.01). Notably, the interindividual differences in RA patients were remarkably large in the third trimester, and they increased postpartum. The expression levels in RA patients were up to 10 times the values of those in healthy controls postpartum, at the time of disease reactivation. The interindividual variation may explain the lack of a significant difference between time points in the RA patient samples for all of the genes but AKT3. AKT3 was the only candidate gene for which significant differences between longitudinally collected samples (third trimester versus postpartum) were found in both groups. Interestingly, RA patients displayed a significantly higher increase in relative AKT3 expression levels from the third trimester to the postpartum time point. Regarding SOCS3 gene expression, we found a variable pattern in healthy donors and in RA patients, with a tendency toward a higher postpartum increase in RA patients (P = 0.051). CD3E showed the highest expression and substantial interindividual variation in expression in both RA patients and healthy donors.

Expression levels in monocytes and T lymphocytes.

Finally, we aimed to define the cellular source of the mRNA for the 5 candidate genes (AKT3, SOCS3, FADS2, CD36, and STAT1). We analyzed gene expression in isolated T lymphocytes (CD3+) and in monocytes (CD14+) in the third-trimester and postpartum samples from both healthy donors and RA patients. Figure 4 shows the data obtained in the third-trimester samples. The data were consistent with the findings in PBMCs (Figure 3). All 5 candidate genes were found to be expressed in both cell populations and to be regulated concordantly. The levels of FADS2 and AKT3 mRNA were minute in monocytes in the third-trimester samples from healthy donors, and the levels of FADS2 were very low in lymphocytes in the third-trimester samples from healthy donors. The postpartum results showed a similar expression profile (Figure 5).

thumbnail image

Figure 4. Gene expression levels in CD3+ and CD14+ cells in the third trimester of pregnancy in rheumatoid arthritis (RA) patients and normal, healthy donors (ND). A, CD3+ and B, CD14+ cells in the samples obtained from healthy donors (n = 7) and RA patients (n = 8) in the third trimester of pregnancy were assessed for mRNA for the indicated genes by quantitative polymerase chain reaction. The relative expression values are shown as the fold change compared to GAPDH mRNA expression. Solid circles represent individual samples; horizontal lines represent the median. ∗∗ = P < 0.01 by Mann-Whitney U test.

Download figure to PowerPoint

thumbnail image

Figure 5. Postpartum gene expression levels in CD3+ and CD14+ cells in rheumatoid arthritis (RA) patients and normal, healthy donors (ND). A, CD3+ and B, CD14+ cells in the samples obtained from healthy donors (n = 7) and RA patients (n = 8) postpartum were assessed for mRNA for the indicated genes by quantitative polymerase chain reaction. The relative expression values are shown as the fold change compared to GAPDH mRNA expression. Solid circles represent individual samples; horizontal lines represent the median. ∗ = P < 0.05 by Mann-Whitney U test.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Human pregnancy represents a unique model for studying the regulation of disease activity in RA. Due to the spontaneous remission that occurs around the third trimester (19, 20), it is possible to identify regulatory factors in the absence of biologic or immunosuppressive agents. One might argue that the decisive immunoregulatory mechanisms that protect the embryo must take place very early in gestation in order to avoid rejection. However, the systemic effect on maternal cells induced by hormonal changes and the transplacental trafficking of fetal antigens increases with advancing gestation (21–23). In other words, PBMCs are clearly exposed to regulatory factors during pregnancy and are likely candidates for influencing arthritic activity in peripheral joints.

Consistent with these considerations, GeneChip experiments using PBMCs in the present study showed an increasing difference in gene expression throughout the trimesters of pregnancy compared to the postpartum time point. Analysis of pathways showed that the most significant differences were in the T cell receptor signaling pathway. This corresponds to our earlier findings in healthy donors and RA patients, which demonstrated a substantial difference between lymphocyte gene profiles in the third trimester and 6 weeks postpartum (9). T cell receptor signaling pathway mRNAs were thus used as internal controls for the subsequent experiments.

Based on the calculated levels of significance and their theoretical potential in immune regulation, we decided to further concentrate on 2 pathways, the adipocytokine pathway and the PPAR signaling pathway. The adipocytokines have recently been shown to modulate inflammation and take part in matrix destruction (24). In addition to their roles in metabolism, they have been found to be expressed in different articular diseases (25–29). The PPAR signaling pathway, which is closely linked to the adipocytokine pathway, mediates antiinflammatory activity in autoimmune disease (30) and plays an important role in the formation and maturation of the placenta (31–33).

Of 19 candidate genes, we decided to further analyze the following 8: SOCS3, STAT1, CD36, PPP3CB, AKT3, CD28, FADS2, and CD3E. The selection of these genes was based on the reproducibility of the differential expression by qPCR and the regulatory potential as deduced from recent studies (9, 34–42). CD28 and CD3E represent the T cell receptor pathway and were used as controls. FADS2 plays a major role during lactation; however, the literature does not indicate that this molecule has regulatory potential in RA (34–36). Thus, it was chosen as an additional control.

The analysis of the prospectively and longitudinally collected samples produced interesting and novel data. First, in the samples from healthy donors, 6 of the 8 candidate genes showed very little interindividual variation, and the expression of 4 of these genes was significantly different during pregnancy compared to after pregnancy. CD28 and CD3E data were consistent with our earlier data, and FADS2 showed the expected regulation. Consistent with their known immunoregulatory roles, AKT3 and PPP3CB expression levels were very low during pregnancy, and STAT1 expression was increased during pregnancy compared to postpartum, in healthy donors (37–39). The data suggest a tight regulation of the expression of these genes, which implies a rather central role of the designated pathways. They may represent established immune changes induced by pregnancy. In contrast, the levels of SOCS3 varied during pregnancy in healthy donors, and there was no uniform decrease postpartum. However, the variation in SOCS3 expression could be explained by its short half-life and its rapid up-regulation in response to proinflammatory and antiinflammatory cytokines, such as interferon-γ, interleukin-1 (IL-1), IL-6, and IL-10 (40, 41). In addition, SOCS3 exerts negative feedback control of cytokines (42). One might therefore hypothesize that greater variations in SOCS3 are needed to stabilize a tolerogenic cytokine profile during pregnancy.

A second finding of the longitudinal analysis was the difference between healthy individuals and those with RA. Notably, the direction of the changes in gene expression during pregnancy was the same in RA patients as in healthy donors. However, the level of expression of 6 of the 8 mRNAs was remarkably higher in RA patients. It appears that, despite clinically defined disease remission in pregnancy, there persists a rather increased, although balanced, molecular activity in PBMCs. Such disconnect between clinical and molecular disease activity has also been found in previously studied cohorts (43). It will be of interest to compare pregnancy-induced disease remission with drug-induced disease remission. To our knowledge, such comparative analyses have not yet been performed.

A third intriguing finding was the large increase in expression levels of several genes in RA patients postpartum. The levels reached up to 10 times the levels in healthy donors and were significantly different (P < 0.01) from those in healthy donors for 6 of 8 mRNAs. It is well established that RA flares weeks to months after delivery (19, 44, 45). Our data strongly suggest reemerging disease activity, and the interindividual variance in expression levels is best explained by different stages of disease reactivation. It remains to be shown to what extent such molecular disease activity precedes a clinical flare.

Based on the longitudinal and comparative data, 2 molecules merit special consideration: SOCS3 and FADS2. Similar to healthy donors, SOCS3 expression showed a mixed pattern in RA patients. SOCS3 is known to be vital for reproduction by promoting placental differentiation and Th2 responses (42). In addition, SOCS3 plays a regulatory role in a murine arthritis model, as evidenced by the fact that its depletion leads to the exacerbation of arthritis (46). Interestingly, SOCS3 was shown to be up-regulated in monocytes in the peripheral blood and synovial fluid of RA patients who were not pregnant (47). A counterregulatory role of SOCS3 in RA might therefore be assumed. FADS2 was initially used as an internal control. Our data, however, clearly showed a difference in FADS2 expression, not only between the pregnancy and postpartum time points, but also between healthy donors and RA patients both in the third trimester and postpartum. The data are comparable to the data for AKT3. Taken together, they suggest an involvement of FADS2 in immune regulation and stress an important role of the adipocytokine pathway in RA.

Recent studies have shown that some of the candidate genes analyzed in the present study are expressed in monocytes and/or lymphocytes (36, 47–50). This is the first study to comparatively analyze gene expression in these cell populations ex vivo. Compared with the findings in PBMCs, the expression levels in these positively selected populations were low. This is likely due to the additional experimental step, which causes a delay in mRNA extraction. Nevertheless, we observed a differential expression that corresponded in quality to the expression found in PBMCs. Again, the expression levels during and after pregnancy were consistent with the anticipated function of the studied molecules.

In summary, our findings suggest a role of the adipocytokine and PPAR signaling pathways in the pregnancy-induced amelioration of RA. We identified several candidate genes, which may be regulatory (SOCS3) or regulated (AKT3, STAT1, FADS2, and CD36). Furthermore, our findings show a substantial reactivation of gene activity around the time of disease flare postpartum.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Villiger 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 conception and design. Weix, Förger, Østensen, Villiger.

Acquisition of data. Weix, Förger, Surbek.

Analysis and interpretation of data. Weix, Förger, Häupl, Villiger.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

We thank Daniel Lottaz, Stefan Wyder, and Stephan Reichenbach for analytical advice, Barbara Helbling, Norina Casutt, and Luigi Raio for help with sample acquisition, and our patients and healthy donors for their blood donation.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ART_34375_sm_SupplFig1.doc45KSupplementary Figure 1

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.