Declaration of Helsinki: The authors state that their study complies with the Declaration of Helsinki that the locally appointed ethics committee has approved the research protocol and that informed consent has been obtained from the subjects.
Correspondence to: D. Bulut, M.D., Division of Cardiology, St. Josef-Hospital, Ruhr-University Bochum, Gudrunstrasse 56, 44791 Bochum, Germany. E-mail: email@example.com
Inflammatory DCM (iDCM) may be related to autoimmune processes. An immunoadsorption (IA) has been reported to improve cardiac hemodynamics. The benefit of IA is probably related to the removal of autoantibodies. A recent study suggests additional effects of IA on the T cell–mediated immune reactions, especially on regulatory T cells (Tregs). In this prospective study, the correlation between the level of Tregs and improvement of myocardial contractility in response to IA in patients with iDCM was investigated. Patients (n = 18) with iDCM, reduced left ventricular (LV) ejection fraction (<35%), were enrolled for IA. Before and 6 months after IA, LV systolic function was assessed by echocardiography, and blood levels of Tregs were quantified by FACS analysis. Patients (n = 12) with chronic ischaemic heart failure and comparable reduced LV-EF served as controls. IA improved LV-EF in 12 of 18 patients at 6-month follow-up. These patients were classified as ‘IA responder’. In 6 patients, LV-EF remained unchanged. At baseline, IA responder and non-responder subgroups showed similar values for C-reactive protein, white blood cells, lymphocytes and T helper cells, but they differ for the number of circulating Tregs (responder: 2.32 ± 1.38% versus non-responder: 4.86 ± 0.28%; P < 0.01). Tregs increased significantly in the IA responders, but remained unchanged in the IA non-responders. In patients with ischaemic cardiomyopathy, none of these values changed over time. A low level of Tregs in patients with chronic iDCM may characterize a subset of patients who do best respond to IA therapy.
Dilated cardiomyopathy (DCM) is defined by an impairment of myocardial contractile function and ventricular dilation. In a subset of patients, the etiopathophysiology of DCM is linked to autoimmune reactions, characterized by the appearance of cardiotoxic autoantibodies in the blood and signs of myocardial inflammation. In about 2/3 of patients with autoantibodies, viral or bacterial RNA or DNA can be detected in myocardial biopsies, suggesting that these immunological features are initiated by an infectious process [1-3]. A (non-ischaemic) DCM with an autoimmune- or immune-mediated infectious background has been termed as inflammatory DCM (iDCM).
A variety of autoantibodies against cardiac cell proteins have been identified in patients with iDCM . Of note, many of these autoantibodies (e.g. targeting ß1-adrenergic receptor, muscarinic M2-acetylcholine receptor, myosin, Na-K-ATPase, troponin I) belong to the IgG subclass 3 that has the highest antibody-dependent potency for cellular toxicity . Wallukat et al.  first reported on the benefit of removal of IgG3 by immunoadsorption (IA) in patients with non-ischaemic DCM, and several open controlled pilot studies confirmed the positive effects of IA on cardiac systolic function, level of natriuretic peptides and quality of life in patients with chronic iDCM or non-ischaemic DCM [6-12]. However, not all studies yielded uniform results [13, 14], and it is until yet unclear which subset of patients with iDCM/non-ischaemic DCM does best benefit from IA therapy. Even if IA is used only once, the level of anti-cardiac antibodies remains low over time . Likewise, a single course of IA treatment shows an increase in left ventricular function over a 6-month period comparable to that after repeated IA treatments at monthly intervals .
A recent study suggests that IA therapy not only removes cardiotoxic autoantibodies from circulation, but also modifies T cell–mediated immune reactions. In this study, IA therapy, which was performed in 10 patients with iDCM, was associated with a significant increase in regulatory T cells (CD4+CD25+CD127low) and a decrease of activated T cells (CD4+/CD69+ and CD8+/CD69+ cells) and CD28+ T cells (co-stimulatory cells) . Regulatory T cells (Tregs; formerly known as T suppressor cells) are important negative immune modulators, constituting of approximately 5% of peripheral CD4+ T cells. They suppress the activation, proliferation and/or differentiation of CD4 and CD8 T cells, B cells, natural killer cells and dendritic cells, thus controlling the immune responses to self-antigens or to pathogens . Depletion or dysfunction of Tregs alone is sufficient to cause autoimmune diseases, vice versa their reconstitution efficiently suppresses autoreactive T cells [16, 17]. Furthermore, Tregs suppress the proliferation of B cells; a depletion of Tregs results in an abnormal humoral response with an increased production of autoantibodies . In mice challenged with coxsackievirus B3, adoptive transfer of Tregs protects against the development of myocarditis by suppressing the immune responses to cardiac tissue .
It is reasonable to assume that changes in T cell regulation and activity in response to IA are linked to inflammatory processes within the myocardium and subsequently myocardial function. In this prospective study, we investigated the correlation between the level of circulating Tregs and improvement of myocardial contractility in response to IA therapy in a consecutive series of patients with iDCM. This study suggests that low levels of Tregs before IA therapy identify a subset of patients who do benefit best from this therapy during a 6-month follow-up.
Subjects and methods
The study population comprises 35 patients recruited in the cardiovascular division of St. Josef-Hospital and BG Kliniken Bergmannsheil, hospitals of the Ruhr-University of Bochum, Germany. Patients (N = 18) were admitted for immunoadsorption. Inclusion criteria were congestive heart failure (CHF) (NYHA II – IV) secondary to chronic iDCM, reduced left ventricular systolic function (EF < 35%), stable medication for CHF for at least 3 months and angiographic absence of coronary artery disease. Patients were stable for ventricular systolic function for at least 3 months. The mean disease duration of iDCM was 14 months, and mean treatment duration, 5 months (range 4–7 months). The diagnosis of iDCM based on previous myocardial biopsies demonstrating immunohistochemical evidence of cardiac inflammation (presence of >14 lymphocytes (CD3+) or macrophages (CD68+)/mm2, diffuse, focal or confluent, enhanced HLA class II expression in antigen-presenting immune cells) according to the World Health Organization/International Society and Federation of Cardiology Task Force on the Definition and Classification of cardiomyopathies , and the absence of cardiotropic viruses (test for human herpesvirus-6, parvovirus B19, Epstein-Barr virus, cytomegalovirus, HIV, ECHO, Coxsackie A/B, Influenza, adenovirus) in cardiac biopsies (as judged by polymerase chain reaction/in situ hybridization). Twelve age-matched patients with chronic ischaemic heart failure and five patients with iDCM who refused IA therapy and with comparable reduced ejection fraction served as controls. Exclusion criteria were clinical or biochemical evidence for the presence of a systemic inflammatory disease, renal insufficiency (serum creatinine >1.8 mg/dl), malignant diseases, thrombocytopenia (<100,000/μl) or anaemia (haemoglobin <11.0 g/dl).
Blood samples were drawn before an IA course of 5 days and 6 months after IA. Before IA treatment and during follow-up visit, clinical examination, routine blood investigations, ECG and transthoracic echocardiography were performed. The echocardiograms Philips iE33 (Philips, Amsterdam, the Netherlands) were performed by cardiologists not related to this study, and unaware of the blood testing results. LV ejection fraction (EF) was derived using Simpson's modified biplane method; left ventricular enddiastolic diameter (LVEDD) was assessed in parasternal longitudinal axis (M-Mode). After insertion of a high-flow catheter into the right jugular vein, 2.5-fold plasma volumes were treated for five consecutive days using protein A agarose columns (Immunosorba; Fresenius Medical Care AG, Bad Homburg, Germany) with acid citrate dextrose solution A (ACD-A) anticoagulation . Plasma was separated for treatment per centrifugation (ComTec; Fresenius Medical Care, Bad Homburg, Germany); protein A agarose columns were inserted in ADAsorb (Medicap, Ulrichstein, Germany) filtration device.
Weight was maintained at a stable level, and furosemide i.v. was applied as necessary. Furthermore calcium carbonate was supplemented orally if patients suffered from paraesthesia or other signs of hypocalcemia. After immunoadsorption, polyclonal immunoglobulin (Intratect®; Biotest AG, Dreieich, Germany) was substituted at 0.5 g per kilogram body weight. In our control patients, (1. with chronic ischaemic heart failure, 2. iDCM who refused IA) blood samples were collected under similar conditions as performed for the patients with iDCM (at baseline and after a 6-month interval).
All study participants gave written informed consent, and the study was approved by the local ethic committee of the Ruhr-University Bochum, Germany.
Lymphocyte and regulatory T cell count by flow cytometry
One hundred micro litres of peripheral blood (EDTA) was incubated for 20 min in the dark with monoclonal antibodies against CD4 (FITC), CD127 (PE), CD25 (APC), CD 3 (PerCP) (all Becton Dickinson, Heidelberg, Germany). Red cells and platelets were lysed (Lysing solution, Becton Dickinson); the remaining mononuclear cells were analysed by flow cytometry (FACs Calibur™; Becton Dickinson) according to standard laboratory guidelines. Each measurement included 50,000 events in the gate of lymphocytes. The numbers of each measured cell population are expressed as per cent of circulating CD3+ T cells if not indicated otherwise. Tregs were characterized as CD4+CD25+CD127low cells. Figure 1 shows a representative FACS analysis.
In a second experiment to quantify Th17 cells, mononuclear cells were prepared by density gradient centrifugation (Polysucrose, Biocoll, Biochrom AG, Berlin, Germany). Further preparing and staining with a cell activation mixture containing the phorbol ester, PMA (Phorbol 12-Myristate 13-Acetate) and a calcium ionophore (ionomycin) were performed according to the manufacturer's guidelines (Leukocyte Activation Cocktail, CD3 (FITC), CD4 (PE), IL-17 (APC), all Becton Dickinson, Heidelberg, Germany). The numbers of each measured cell population are expressed as per cent of circulating CD3+ T cells if not indicated otherwise. Th17 cells were characterized as CD3+CD4+IL17+ cells.
Data are expressed as mean + SEM, if not indicated otherwise. Differences in T cell counts between the groups were examined by student's t-test; differences in course of time were examined by anova followed by Bonferroni post hoc test. Statistical significance was assumed at P < 0.05.
Twelve of 18 patients (67%) with iDCM who underwent IA therapy showed an improvement of left ventricular systolic function (defined as an increase in EF > 5%) at 6-month follow-up as compared to the corresponding values before IA. These patients were defined as ‘IA responder’. In 6 patients, left ventricular ejection fraction remained almost unchanged during a 6-month observation period (Δ EF ≤ 5%): these patients were defined as ‘IA non-responder’. Table 1 summarizes the baseline characteristics and demographic data for the patients with iDCM (all, IA responder, IA non-responder) before and 6 months after IA therapy. For comparison, the data from 12 patients with chronic ischaemic cardiomyopathy and 5 patients with iDCM without IA at the time of inclusion and 6 months later were also given. Of note, left ventricular ejection fraction and enddiastolic diameter improved significantly in the IA responder group only. Not unexpected, left ventricular indices were unchanged in the control group (ischaemic heart failure). The groups did not differ with respect to mean age, body mass index, systolic and diastolic blood pressure and co-medication for heart failure.
Table 1. Baseline characteristics, demographic data
Table 2 summarizes the laboratory findings. At baseline, the IA responder and non-responder subgroups showed similar values for C-reactive protein (CRP), white blood cell count, lymphocyte count and CD4+ T helper cells, but they differ significantly for the number of circulating Tregs (responder: 2.32 ± 1.38% versus non-responder: 4.86 ± 0.28%; P < 0.01). Six months after IA therapy, the values for CRP, white blood cell count, lymphocyte count and CD4+ helper T cells remained almost identical for the IA responder and IA non-responder subgroups. Tregs increased significantly in the IA responder subgroup by on average 75%, but remained unchanged in the IA non-responder subgroup. In patients with ischaemic cardiomyopathy, none of these values changed over time (6 months) significantly (Table 2).
Table 2. Effect of IA on Leucocyte and Lymphocyte Counts for all patients underwent IA
Figure 2 demonstrates the Treg values for individual patients before IA therapy. Please note that all 12 patients with iDCM who experienced an improvement of LV systolic function after IA therapy had at baseline low Tregs <4%, whereas the 6 non-responders had Tregs ≥4% at baseline. The improvement of ejection fraction correlated positively with the raise in Treg count (r = 0.62). Figure3 illustrates the number of Tregs before and 6 months after IA for responder and non-responder.
In addition to these results, responding and non-responding patients differ significantly in the number of Th17-cells (responder: 1.41 + 0.33% versus non-responder: 0.71 ± 0.26%; P < 0.01). After IA treatment, Th17-cells decreased significantly in the IA responder subgroup, but remained unchanged in the IA non-responder subgroup (Table 2).
Viral proliferation in cardiac tissue and the host immune response to eliminate the virus characterize the pathogenesis of viral myocarditis. This host immune response is accompanied by autoimmune and/or autoreactive processes, related to a molecular mimicry between viral and host antigenic epitopes or to epitopes exposed by injured cardiomyocytes. All three events (virus infiltration of cardiomyocytes, immune cells targeting virus-infected cardiomyocytes and production of circulating autoantibodies and/or autoreactive immune cells) are discussed to participate in the destruction of cardiomyocytes . Even after elimination of the virus, autoimmune processes may still be ongoing, finally leading to dilated cardiomyopathy.
The patients of the present study, who were enrolled for the IA therapy, are suffering from non-ischaemic DCM. They are characterized by the immunohistochemical evidence of cardiac inflammation and absence of cardiotropic virus genome, and were classified according to the WHO criteria  as patients with iDCM. In 1996, Wallukat and coworkers reported on the benefit of removal of autoantibodies to the ß1-adrenergic receptor by IA in 8 patients with non-ischaemic DCM. As the autoantibody titre decreased, the systolic cardiac function and symptoms improved. Subsequently, this technique has been applied to more than 200 patients with non-ischaemic DCM/iDCM [6-12]. Most groups used columns containing sheep anti-IgG antibody and protein A agarose for the removal of autoantibodies. The effect of protein A agarose column immunoadsorption is non-specific, removing pathologic and non-pathologic antibodies . Because of the concern of humoral immunodeficiency due to this non-specific removal, most groups (including ours) supplement immunoglobulin after completion of IA therapy. A Japanese group used a tryptophan column with a high specificity for the IgG-3 subclass in 16 patients with non-ischaemic DCM and noted a significant decrease in plasma B-type natriuretic peptide . The authors recommend that the selective removal of IgG-3 does not necessitate immunoglobulin supplementation.
Not all studies using the IA therapy in patients with non-ischaemic DCM yielded uniform results. Cooper and coworkers  pointed out that the response to IA treatment on regional LV function is not uniform, although the quality of life assessment yielded significant improvement up to 6 months after IA in patients with chronic DCM. Doesch and coworkers  performed IA in a series of 27 patients with chronic non-familial DCM and reported on an improvement of markers of heart failure severity (NT-pro-BNP) in a subset of patients only, with a non-significant improvement in LV systolic function indices. Also, in the present study, we noted an improvement of systolic LV function (as defined as an increase in LV ejection fraction >5% after a 6 months observation period) in only 67% of patients with iDCM, and in the remaining patients, the systolic LV function remained almost unchanged (at least during a 6-month follow-up). This obvious discrepancy between the published studies may be related to several aspects, among others the lack of standardized selection criteria for the underlying cardiac disease (chronic non-familial DCM, non-ischaemic DCM, idiopathic DCM, chronic DCM, iDCM, healed iDCM), and differences in the definition of ‘benefit of IA therapy’. Furthermore, a randomized, prospective study is still lacking; thus, the scientific evidence for a beneficial effect of removal of IgG-3 from blood circulation (versus an adequate control group) on cardiac function is unknown. Autoantibodies belonging to IgG3 group may play a pivotal role in cardiac dysfunction of patients with iDCM. Staudt et al.  could show in a case–control study that Protein A agarose column immunoadsorption in conjunction with an improved treatment regimen for IgG3 elimination induces hemodynamic benefit in patients suffering from DCM.
The present study focuses on Tregs in response to IA therapy and intravenous IgG substitution. There is little doubt that the cell-mediated immunity is a key player in the pathogenesis of myocarditis and post-inflammatory cardiomyopathy. In knockout mice, elimination of both CD4+ and CD8+ T cells protected the animals from coxsackie B3 viral myocarditis. Likewise, mice with a deletion of alpha-beta T lymphocytes had only minimal myocardial damage after coxsackie viral infection . Regarding Tregs, numerous studies reported decreased levels of Tregs and/or suppressed Treg function in patients with myocarditis or idiopathic cardiomyopathies [25-29]. In the present study, similar blood levels of Tregs (defined as CD4+CD25+CD127low and expressed as% CD3+ T cells) were observed in patients with iDCM and age-matched patients with stable and chronic ischaemic cardiomyopathy. A novel finding is that iDCM patients with low levels of Tregs (<4%) showed a significant of improvement of systolic LV function after IA therapy, whereas patients with higher levels (≥4%) did not respond to this treatment. The number of Tregs increased in responders in the observation period and shows no difference to other groups 6 months after IA.
In addition to these results, we found that another subset of helper T cells is influenced by IA + IgG substitution. These Th17 cells play an important role in the induction of autoimmune tissue injury. They are distinct from Th1 or Th2 cells because they do not produce classical Th1 or Th2 cytokines such as IFN-γ or IL-4. There is a functional antagonism between Th17 and Treg cells. Both populations are regulated by variable levels of TGF-ß and IL-6. At a steady-state level or in absence of inflammatory stimuli, TGF-ß suppresses the generation of T effector cells and induces FoxP3 regulatory T cells and thereby maintain self-tolerance. In state of inflammation, IL-6 suppresses the generation of TGF-ß-induced Treg cells and induces a pro-inflammatory T cell response predominated by Th17 cells [30, 31].
In our study, IA-responding patients had higher levels of Th17 cells compared to non-responders and control patients with ischaemic heart failure. These observations have to be confirmed in larger trials. But this observation may be a first step to characterize a subgroup of patients with iDCM who do best benefit from IA therapy. It is not known how IA therapy can affect cell-mediated immune responses. Particularly, it is not known whether non-specific removal of IgG antibodies and/or non-specific ‘immune-modulatory’ effects secondary to plasmapheresis and/or IgG substitution after IA are responsible for this phenomenon. Autoantibody-induced inflammation can be separated into two components, autoantibody production and local inflammatory response. Tregs suppress both components, thereby controlling autoimmune inflammation. Follicular Tregs may suppress follicular T helper cell–mediated antibody production. CD4+CD25+FoxP3+ Tregs have the capacity to control inflammation by suppressing cytokine production in T helper cells. Furthermore Tregs are able to suppress innate cells via IL-10 production. These IL-10 producing cells may also play a pivotal role in regulating Th17 cells .
On the basis of our results, it appears, however, reasonable to integrate the measurement of T cell–meditated immune responses, in particular Tregs levels, in future trials on the effects of IA therapy on LV function in patients with iDCM.
The screening process and inclusion criteria were quite restrictive. Thus, the sample size of our study is small, and this may limit our conclusions. Furthermore, an appropriate control group is lacking who underwent ‘sham - immunoadsorption therapy’. In our small control group of patients who refused IA therapy, we postulated to examine changes in cellular immunity during progression of the disease, but we cannot verify this topic. We cannot rule out confounding (and yet unknown) factors that might have influenced cell-mediated immunity and benefit of IA. Furthermore, we did not analyse the auto-antibody status in our patients. So we cannot rule out confounding factors that (1) antibodies' levels may influence our results and (2) patients with ischaemic cardiomyopathy may have auto-antibodies against myocardial targets. We did not examine whether patients with ischaemic cardiomyopathy would benefit of IA too. IA was performed as described previously by several investigators [5, 6, 12]. In these protocols, IA was followed by substitution of polyclonal immunoglobulins. We cannot disclose confounding factors of IG substitution, which may interact with cellular immunity.
Different ways are known to analyse Tregs. Tregs are broadly classified into natural Tregs (CD4+CD25+), which emigrate from the thymus to perform the key role in immune homoeostasis, or adaptive Tregs (non-regulatory CD4+ T cells),which acquire CD25 expression outside of the thymus. They are typically induced by autoimmunity . Recently, the transcription factor forkhead box p3 (FOXP3) has been reported to play a major role in CD4+ CD25+ Treg function and represents a specific marker for these cells. However, FOXP3 is a nuclear protein and is of limited value in the isolation of Tregs, which is a major reason that many functionally relevant aspects of Treg cells are still unknown . In this work, we did not analyse FOXP3. In addition to cellular aspects, we did not analyse genetic polymorphisms in Fcγ-Receptor IIa as it was described previously .
Conflict of Interest
This work was supported by a research grant from Fresenius Medical Care, Bad Homburg, Germany.
Our group examined for the first time to our knowledge T cell subgroups in immunoadsorption in patients with dilated cardiomyopathy . The actual study population was recruited after the publication of above-mentioned work. So none of the patients in this work was included in the previous work.