Functional autoantibodies in systemic sclerosis: influence of autologous stem cell transplantation and correlation with clinical outcome

Abstract

Objectives

To evaluate the effect of autologous stem cell transplantation (aSCT) on functional antibodies (abs) to the angiotensin II type-1-receptor (AT₁R) and topoisomerase-I (topo-I) in SSc-patients and to analyse their prognostic relevance.

Material and methods

Forty-three SSc-patients in whom aSCT was performed were analysed. Thirty-one patients had a favourable outcome after aSCT (group 1), 12 patients showed no response or relapse (group 2). Patients’ sera were tested for anti-AT₁R and anti-topo-I antibodies by ELISA and in a luminometric assay (LA) using AT₁R-expressing Huh7-cells for inhibitory or stimulatory anti-AT₁R antibodies before and after aSCT (4–217 months, median 28 months). Anti-topo-I antibodies were also analysed for their capacity to inhibit enzyme function.

Results

A total of 70% of the SSc patients had anti-topo-I- and 51% anti-AT₁R antibodies in the ELISA before aSCT. In all instances, anti-topo-I antibodies inhibited topo-I-enzyme function. In the LA, 40% had stimulatory and 12% inhibitory anti-AT₁R antibodies. Anti-topo-I- and anti-AT₁R-reactivity (ELISA) significantly decreased after aSCT. Before aSCT, anti-topo-I-reactivity was significantly higher in group 2 patients than in group 1 patients (P < 0.001), while there was no difference between both groups for anti-AT₁R antibodies detected by ELISA. Stimulatory anti-AT₁R antibodies detected by LA were confined to group 1-patients.

Conclusions

Reactivity of functionally active anti-AT₁R antibodies was not influenced by aSCT, while anti-topo-I antibodies decreased after aSCT. The fact that anti-topo-I antibodies inhibited enzyme function in all instances supports the hypothesis of a pathogenetic role of the topo-I antigen/antibody-system in SSc. High anti-topo-I reactivity before aSCT was associated with an unfavourable, presence of stimulatory anti-AT1R antibodies with a favourable course after aSCT.

Introduction

SSc is an autoimmune connective tissue disease characterized by inflammation and fibrosis involving skin, the vascular system and internal organs, especially the lung [1]. Its pathogenesis is unknown; genetic, environmental, and immunological factors and mechanisms mediating cell senescence or inflamm-aging are postulated to be involved [1–3]. The latter factors include oxidative stress, chronic inflammation, vascular insufficiency, DNA damage, telomer shortening and autoantibody formation [2, 4, 5]. Thus, antinuclear antibodies have been particularly described as a hallmark of SSc as, for instance, antibodies to DNA topoisomerase-I (topo-I; Scl70), centromeres or nucleoli (fibrillarin) [6, 7]. Anti-topo-I antibodies are present in 20–30% of SSc-patients and preferentially associated with diffuse cutaneous involvement and pulmonary interstitial fibrosis indicating an unfavourable course [8, 9]. Their pathogenic relevance is controversially discussed, but they are useful routine markers for diagnosis and prognosis of organ involvement. Interestingly, it has been shown in early studies by assaying the relaxation of superhelical DNA that anti-topo-I antibodies inhibit topoisomerase-I activity [10]. They may, therefore, belong to the group of functional antibodies because autoantibodies are considered ‘functional’ when they directly activate or inhibit molecular pathways by binding to the identified target antigen [11].

Meanwhile, a variety of other functional antibodies reacting in particular with receptors on cell membranes have been described in SSc. Thus, for instance antibodies to angiotensin (AT₁R) and endothelin receptors (ET_A1) have been hypothesized to play a pathogenic role in SSc although they have been observed in several other disorders [12–14]. First studies on anti-AT₁R- and -ET_A1 antibodies were based on true functional bioassays that involved spontaneously beating cultured rat cardiomyocytes or human endothelial cells [15, 16]. However, those assays are time consuming and difficult to standardize for routine use. Therefore, solid phase assays were established with extracts from Chinese hamster ovary (CHO)-cells overexpressing the AT₁R or ET_A1 [16, 17]. They are commercially available and have been applied in several studies [12–14]; indeed, immunoglobulins containing antibodies measured by these ELISAs may influence the receptor functions [12, 13, 18]. Nevertheless, it is still a matter of debate whether all these antibodies are really functionally active according to the aforementioned definition because binding to a receptor does not necessarily reflect signalling. Moreover, we could show that fractions containing cell membranes overexpressing the receptors as used in these ELISAs can still be contaminated with slight amounts of further nuclear or cytoplasmic antigens leading to a reaction with other autoantibodies present in patients’ sera [14].

In a recent study we established a true functional assay that allowed the reliable and reproducible detection of functional anti-AT₁R antibodies using Huh7-cells overexpressing the receptor [14]. Antibodies inhibiting or stimulating the AT₁R-function were observed in 52% of the SSc patients, but they did not correlate with disease activity or clinical manifestations. Moreover, they were also observed in patients with other disorders and did not correlate with the anti-AT₁R antibodies determined by ELISA [14].

There are hardly any therapeutic options for general improvement or termination of fibrosis in SSc [19]. Autologous hematopoietic stem cell transplantation (aSCT) was introduced for SSc treatment in 1997 and has since then been shown to be the most effective therapeutic option for selected patients with severe and refractory disease [20–22]. It has been hypothesized that aSCT in autoimmune disorders leads to an ‘immune reset’ by eradication of auto-reactive lymphocytes by immunoablative conditioning, and/or the correction of dysregulated immune balance by newly developed (regulatory?) lymphocytes derived from transplanted hematopoietic stem cells [23]. In a recent study we showed that anti-topo-I-reactivity decreased after aSCT, but only in a few patients it became negative [24].

The influence of aSCT on functional autoantibodies has not yet been analysed in SSc. The aim of the present study was, therefore, to see whether functional anti-AT₁R antibodies measured by different assays, but also the ability of anti-topo-I antibodies to inhibit enzyme function is influenced by aSCT and whether there may be a correlation with the clinical course.

Patients

Forty-three SSc patients in whom aSCT had been performed (25 females, 18 males) were analysed. Diagnosis was in accordance to the 2013 Classification Criteria for SSc [25]. Clinical data from 32 of these patients have been presented in previous studies [20, 22, 24]. In the meantime (since 2012), a further 11 patients have received aSCT. Only patients with a follow up of >4 months after aSCT were included in the study (median 28 months, range 4–217 months). Clinical data of the 43 patients as well as inclusion and exclusion criteria and transplant regimen are given in Supplementary Data S1, available at Rheumatology online.

The 43 patients were classified according to the outcome after aSCT as patients with good response and no relapse (group 1, n = 31) vs those with primary response but relapse within the first 24 months after aSCT or without response (group 2, n = 12) during a long-term follow up as recently described [22] (see also Supplementary Data S1, available at Rheumatology online).

From all 43 patients, sera were available before transplantation [time point (tp) 0]. Six time points after aSCT were chosen for serological analysis: tp1: 1–4 months (n = 39); tp2: 5–9 months (n = 34); tp3: 10–17 months (n = 27); tp4: 18–24 months (n = 24); tp5: 25–36 months (n = 20); and tp 6: >36 months after aSCT (n = 11). In total, 198 serum samples were analysed. Blood was drawn during routine testing. All patients had been regularly seen by at least one of the authors (J.H., A.-C.P.). Moreover, sera from 15 healthy controls were included (derived from laboratory staff). Patients’ sera were stored at –20°.

The study had been approved by the local ethical committee of the University of Tuebingen (No.681/2011BO2; 647/2016BO2); it was performed according to the Helsinki guidelines, and patients had given written informed consent before the study.

Methods

Isolation of immunoglobulins

In functional assays it is mandatory to use purified immunoglobulins instead of whole sera in order to exclude the effect of other stimulatory or inhibitory factors that may be present in sera.

Immunoglobulins were purified from patients’ sera by ammonium sulphate precipitation [26]. The IgG-concentration within these immunoglobulin fractions were randomly determined revealing concentrations of about 10 µg/µl.

Immunoglobulins could be purified from 25 patients before aSCT (tp0) and at tp2 (n = 25), tp3 (n = 20) and tp4 after aSCT (n = 17; in total n = 87). These 25 patients were representative for the whole cohort with respect to clinical and epidemiological data and outcome after aSCT. At tp1 not enough amount of serum was available from most patients to isolate immunoglobulins. At later time points (tp5, 6) number of patients was too low for statistical analysis. Therefore, for these experiments with purified immunoglobulins only the tp0, 2, 3 and 4 were analysed.

Methods for detection of autoantibodies

Anti-topo-I antibodies were determined by a published in-house assay using a recombinant full-length topoisomerase-I obtained from Diarect (Freiburg, Germany) [24]. Positive and negative standard sera were used in each test to calculate a standard curve. Results are given as absorbance×$$1.000. Cut-off values have been determined as described [14]; for the anti-topo-I-ELISA they had been determined in previous studies with large numbers of healthy individuals. Mean of absorbance (×1.000)+3-fold standard deviation resulted in a cut-off >300 that was also verified by receiver operating curves (ROC). It was re-analysed in the present study with sera from 15 healthy individuals and found to be identical.

Anti-AT₁R antibodies were measured by ELISA with an assay from CellTrend GmbH (Luckenwalde, Germany). Results are given as U/ml. The normal value of the anti-AT₁R-assay given by the manufacturer (17 U/ml) was re-analysed and confirmed with sera from healthy individuals.

The bioassay for the demonstration of functionally active anti-AT₁R antibodies has been performed as recently described using Huh7-cells constitutively expressing the AT₁R. Cells were transfected with an aequorin/green fluorescence protein fusion plasmid [14]. Ammonium sulphate precipitated immunoglobulins from patients’ sera (0,25  µg/µl) were added to the cells for 1 h. The reaction was started with the injection of 10 µM of the AT₁R agonist angiotensin II (Merck, Darmstadt, Germany). The change in intracellular (Ca⁺⁺) during 20 s was then determined by measuring the emitted light with a 2460 MicroBeta²LumiJET-luminometer (Perkin Elmer, Downers Grove, IL, USA). Measurements were performed in quadruplicate. Results are given as a factor: relative light units (RLU) were measured with each immunoglobulin fraction and given as percentage of RLUs without added immunoglobulins (%RLU); in order to determine a ‘normal range’, the immunoglobulins from 15 healthy individuals were analysed, and the mean of the obtained %RLU was calculated. %RLU from patients and controls were divided by this mean resulting in a factor. A factor <0.6 was defined as inhibitory, a factor >1.4 as stimulatory activity.

Topoisomerase-I-inhibition assay

Topo-I-enzyme activity was monitored by assaying the relaxation of superhelical DNA [27] as described with some modifications [10]. Recombinant topo-I (2.8 µg/ml) was incubated with purified immunoglobulins (1:100) for 1 h at room temperature. Supercoiled pUC19 plasmid DNA (10 ng/ml) was added for 30 min at 37°C. The effect of the topoisomerase on the plasmid DNA was visualized using standard agarose gel electrophoresis.

All 87 immunoglobulin fractions were analysed for their capacity to inhibit topo-I-enzyme-activity. Results are given as positive or negative (inhibition or no inhibition of topo-I by the immunoglobulin fractions).

Statistics

For statistical analysis, SPSS version 15.0 and GraphPad Prism7 were used. Paired data were analysed by Wilcoxon test, unpaired data by Mann–Whitney U-tests. Fisher’s exact test was used for comparing prevalences. Correlation was evaluated by the Spearman Rank test. Values of P < 0.05 were considered statistically significant.

Kaplan–Meier curves comparing autoantibody reactivity in group 1 and group 2 patients before aSCT (tp0) and the corresponding area under the curves (AUC) were calculated by GraphPad Prism7.

Results

Influence of aSCT on antibody reactivity

Anti-topo-I- and -AT₁R antibodies measured by ELISA

Anti-topo-I- and anti-AT₁R antibody reactivity significantly decreased since tp1 until the end of the observation period (tp6) and until tp4, respectively (Fig. 1a). Also, the prevalence of both antibodies decreased, but the difference was statistically significant only for the anti-topo-I antibodies at tp3-5 (Table 1a); i.e. the effect of aSCT was stronger on the production of anti-topo-I- than anti-AT₁R antibodies. Nevertheless, there were patients in whom anti-topo-I antibodies persisted for 10 years and longer with high reactivity.

Anti-AT₁R antibodies measured in the functional assay

The functional assay was performed with immunoglobulins from 25 of the SSc patients before transplantation (tp0) and at tp 2–4 (in total: 87 sera).

Stimulatory antibodies predominated (40%) as compared with inhibitory antibodies (12%) (Table 1b). Prevalence and reactivity of these antibodies did not change after aSCT (Table 1b, Fig. 1b). The ratio between stimulatory and inhibitory antibodies decreased from 3.3 (tp0) to 1.4 (tp4).

Functional anti-topo-1-antibodies

Of the 25 patients analysed, 16 were anti-topo-I-positive by ELISA; 15 of them (94%) inhibited topoisomerase-I activity (Table 1c); this effect was not observed with any of the immunoglobulins from anti-topo-I-negative sera; a representative picture for an anti-topo-I-positive and -negative serum is given in Fig. 2. After aSCT, the antibodies retained their functionality in all instances (Table 1c).

Comparison of antibody reactivities measured by the different methods

Comparing the anti-AT₁R-ELISA with the functional AT₁R-assay in patients before aSCT, there was no correlation between both methods (Spearman correlation coefficient r=–0.11).

There was a strong correlation between anti-topo-I- and anti-AT₁R-reactivity measured by ELISA (r = 0.715) and between anti-topo-I antibodies measured by ELISA and relaxation assay (r = 0.91; Table 1c).

Antibody reactivity in relation to clinical outcome after aSCT

Anti-topo-I- and -AT₁R antibodies measured by ELISA

Patients of group 2 showing relapse or no response after aSCT had significantly higher anti-topo-I antibody reactivity before aSCT (tp0) than patients of group 1 (P < 0.01; Fig. 3a). In both groups antibody reactivity significantly decreased after aSCT. It was still significantly higher at tp1 in group 2 than in group 1 (P < 0.05), but in the follow-up there were no differences any more between both groups. The prevalence of the anti-topo-I antibodies was also higher in group 2 patients (92%) than in group 1 patients (61%) at tp0, but the difference was not statistically significant (P = 0.07; Table 2a). In group 2, prevalence decreased after aSCT without reaching statistical significance. Comparing the prevalence between both groups at the different time points after aSCT, there were no significant differences (Table 2a).

In contrast, there was no significant difference in anti-AT₁R-reactivity (ELISA) between group 1 or 2 patients before aSCT (Fig. 3b). Anti-AT₁R antibodies decreased in both groups after aSCT, but to a less extent than the anti-topo-I antibodies. Reactivity did not differ significantly between both groups at any tp after aSCT. Prevalence of anti-AT₁R antibodies did not significantly change after aSCT and did not differ between group 1 and group 2 patients (Table 2a).

These data were also confirmed calculating the Kaplan–Meier curves at tp0 comparing group 1 and 2 patients. AUC calculated in group 1 and group 2 patients significantly differed for anti-topo-I (P = 0.02) but not for anti-AT₁R antibodies (P = 0.31; Fig. 4a, b).

Anti-AT₁R antibodies measured in the functional assay

At tp0, none of the group 2 patients had stimulatory anti-AT₁R antibodies in contrast to 53% of group 1 patients (trend with P = 0.051; Table 2b). Also, reactivity was lower in group 2 than in group 1 patients (trend with P = 0.065; Fig. 3c). Inhibitory antibodies were found in 11% of group 1 and 17% of group 2 patients; their prevalence increased from tp0 to tp4 (Table 2b), but the difference was not statistically significant. The ratio between stimulatory and inhibitory antibodies in group 1 gradually decreased from 5.0 (tp0) to 1.75 (tp4). After aSCT, the median of the functional activity decreased in both groups but the difference was statistically not significant (Fig. 3c).

Again, these data could be confirmed by Kaplan–Meier curves calculated at tp0 revealing a significant difference in the AUC between group 1 and group 2 (P = 0.01, Fig. 4c).

Functional antibodies to topo-1

Similar to the anti-topo-I antibodies determined by ELISA, the prevalence of those detected by topoisomerase-relaxation test was also higher in group 2 than in group 1 patients; in both groups, the number decreased after aSCT without reaching statistical significance due to low patient numbers (data not shown).

Discussion

In this study, analysing a large number of SSc patients after aSCT for >36 months we showed that the reactivity of inhibitory or stimulatory antibodies to the AT₁R measured by a recently described functional assay using AT₁R expressing Huh7 cells [14] was not affected by autologous stem cell transplantation. Also, anti-AT₁R antibodies determined by a commercially available ELISA-kit only slightly decreased after aSCT. Moreover, the antibodies did not correlate with prognosis after aSCT. This is in contrast to the anti-topo-I antibodies which significantly decreased after transplantation—although they did not become completely negative—and were found in higher prevalence and reactivity in patients with relapse or no response after aSCT than in patients with good prognosis, hereby confirming our previous data [24]. Interestingly, these antibodies are in nearly all instances functionally active, i.e. they inhibited topo-I-enzyme activity in a DNA relaxation test as already reported in the 1980s [10].

In our previous study we showed that anti-AT₁R antibodies detected by ELISA correlated—similar to the topo-I antibodies—with some clinical symptoms such as digital ulcers, pulmonary fibrosis and oesophageal manifestations. In contrast, the functionally active anti-AT₁R antibodies measured by luminometric assay were not associated with distinct clinical symptoms [14]. From the present study, however, there is strong evidence that the presence of stimulatory anti-AT₁R antibodies may be indicative for an unfavourable course after aSCT. Interestingly it has been shown in an animal model that anti-AT₁R antibodies induced in mice as well as injection of monoclonal anti-AT₁R antibodies lead to SSc-like symptoms such as skin and lung inflammation or lymphocytic alveolitis [28]. In vitro it has been shown that anti-AT₁R antibodies induce ERK 1/2 phosphorylation and increase expression of transforming growth factor beta messenger RNA, vascular cell adhesion molecule 1 and interleukin 8 in endothelial cells [13]. Moreover, it is known that neutrophils express the AT₁R, and it has been, therefore, argued that the antibodies may activate neutrophil AT₁R within blood vessels and that these activated neutrophils are home to areas of inflammation and exacerbate tissue damage [29]. These data fit to our observation that in SSc predominantly stimulatory antibodies were found and that the ratio of stimulatory to inhibitory antibodies decreased in patients with good prognosis after aSCT. This may, indeed, indicate a clinical relevance for the stimulatory antibodies; but this has to be proven in further studies.

In contrast to the functional anti-AT₁R antibodies, anti-topo-I antibodies show a strong disease specificity and correlate with different clinical manifestations [8, 9, 14], strongly suggesting a pathogenic role. Interestingly, topotecan, an inhibitor of topoisomerase-I used in cancer therapy has been shown to induce SSc-like disease [30]. Moreover, it was shown that topoisomerase-I binds specifically to fibroblasts; this complex was recognized by anti-topo-I antibodies, induced cytokine-like effects, stimulated fibroblast migration and promoted fibrosis [31–33]; i.e. anti-topo-I antibodies may, indeed, influence fibroblast function in SSc [33, 34]. Autoimmune properties of B cells are increased in SSc patients as shown by an overexpression of CD19 and B-cell activating factor [35, 36]. However, T cells also play an important role because production of anti-topo-I antibodies by autologous B cells obtained from SSc patients depends on interleukin-6 secretion by topo-I-specific T-cell clones [37], and interstitial fibrosis seems to be promoted by disturbed T-cell subsets and by secretion of pro-fibrotic cytokines from CD8+ T cells [38, 39]. Also, studies in mice have shown that induction of an autoimmune response to topo-I causes skin and lung fibrosis characteristic of SSc, and that besides B-cells and autoantibodies, T cells also play an important role [40, 41].

Introduction of aSCT as treatment for progressive SSc and other autoimmune disorders was based on the concept that it may lead to an ‘immune reset’ by eradication of auto-reactive lymphocytes by immunoablative conditioning, and/or the correction of dysregulated immune balance by newly developed lymphocytes derived from transplanted haematopoietic stem cells [23]. Immune reconstitution after aSCT has been analysed only in small cohorts of SSc patients. Data are still inconclusive. A sustained reduction of CD4+ T cells and B cells, suppression of serum autoantibody and cytokine concentrations, and a shift from Th2 to Th1 cells have been reported [23]. However, an increased total B-cell percentage within the lymphocytes was also seen [42], which was related to increased naive B cells. Anti-topo-I antibodies remained abnormally high throughout 24 months after aSCT [42] and even longer, as shown in the present study. These data may indicate either that aSCT did not lead to re-programming the immune system at least in these patients, that it was inefficient to eliminate the autoreactive cells including long-living plasma cells still producing autoantibodies or that the CD34+ progenitor cells reinfused may be prone to differentiate again into autoreactive cells directed against topo-I. Plasma cells can survive in bone marrow niches despite cytoreductive or immunosuppressive therapies, which could explain refractory disease or relapse [43, 44].

Our study has several limitations. Thus, number of patients before and after aSCT of whom immunoglobulins could be isolated was rather low due to low serum amounts, which may affect the statistical power. One has also to be aware that the determination of the prevalence of an antibody always depends upon the definition of cut-off values. For the bioassay, a factor of <0.6 and >1.4 for inhibitory and stimulatory antibodies, respectively, was accepted which resulted in at most 15% of healthy controls showing either inhibitory or stimulatory antibodies. However, there may exist inhibitory, stimulatory and neutral antibodies to the same receptor in one serum reacting with different epitopes [45], i.e. in patients’ sera we can always detect only the predominating form of antibodies. Immunoglobulins from patients showing in our functional assay ‘no effect’ on the receptor may contain both stimulatory and inhibitory antibodies in similar concentrations; i.e. we can also not exclude that healthy individuals contain functionally active anti-AT₁R antibodies and that the stimulatory and inhibitory antibodies balance out one another [46]. The predominance of one type of antibodies may then be indicative for an immunological dysbalance, but whether this is correlated with clinical symptoms still remains obscure. Moreover, we used Huh7-cells for the functional assay which is a human cell line derived from hepatic carcinoma; hepatocytes are not the major target of autoimmunity in systemic sclerosis and our system is, therefore, somewhat artificial and may not reflect the reality. However, this may hold true also for CHO cells overexpressing the AT₁R receptor, which have been used in other studies [16, 47]. We found similar data comparing a functional assay with CHO- and Huh7 cells [14]. In contrast, endothelial cells from umbilical cord were not suitable for our functional assay because they did not express the AT₁R (own unpublished observation). Of course, it cannot be excluded that also functional autoantibodies reacting with a receptor expressed by different tissues may have some organ or disease-specific function by binding probably to a receptor in combination with tissue-specific proteins required for a distinct signalling, i.e. anti-AT₁R antibodies in different disorders may recognize different molecular forms of human AT₁R.

In conclusion, we have shown that despite the clinically well-documented benefit of aSCT for patients with severe SSc, reactivity of functional anti-AT₁R antibodies was not influenced by this kind of therapy. In contrast, anti-topo-I antibodies that could also be proven to be functional active in inhibiting enzyme function significantly decreased, although they did not become negative. These data are, again, strongly suggestive of a role of the topo-I antigen/antibody system in the pathogenesis of SSc. Interestingly, high anti-topo-I reactivity before aSCT was significantly associated with an unfavourable and presence of stimulatory anti-AT₁R antibodies with a favourable course after aSCT, indicating that these antibodies may serve as biomarkers for the clinical response to this kind of therapy.

Supplementary data

Supplementary data are available at Rheumatology online.

Data availability statement

The data underlying this article will be shared on reasonable request to the corresponding author.

Funding

No specific funding was received from any bodies in the public, commercial or not-for profit sectors to carry out the work described in this article.

Disclosure statement: The authors have declared no conflicts of interest.

Acknowledgements

The study had been approved by the local ethical committee of the Medical Faculty and University Hospital of the Eberhard Karls University, Tuebingen (No. 681/2011BO2; 647/2016BO2); approval was for the analysis of humoral and cellular immunological reactions in patients with chronic rheumatic disorders using stored blood samples. The study was performed according to the Helsinki guidelines, and patients had given written informed consent before the study.

References

Author notes