Blood-based microRNA profiling in patients with cardiac amyloidosis

Introduction Amyloidosis is caused by dysregulation of protein folding resulting in systemic or organ specific amyloid aggregation. When affecting the heart, amyloidosis can cause severe heart failure, which is associated with a high morbidity and mortality. Different subtypes of cardiac amyloidosis exist e.g. transthyretin cardiac amyloidosis and senile cardiac amyloidosis. Today, diagnostics is primarily based on cardiac biopsies and no clinically used circulating blood-based biomarkers existing. Therefore, our aim was to identify circulating microRNAs in patients with different forms of amyloidosis. Methods Blood was collected from healthy subjects (n = 10), patients with reduced ejection fraction (EF < 35%; n = 10), patients affected by transthyretin cardiac amyloidosis (n = 13) as well as senile cardiac amyloidosis (n = 11). After performing TaqMan array profiling, promising candidates, in particular miR-99a-5p, miR-122-5p, miR-27a-3p, miR-221-3p, miR-1180-3p, miR-155-5p, miR-339-3p, miR-574-3p, miR-342-3p and miR-329-3p were validated via quantitative real time PCR. Results The validation experiments revealed a significant upregulation of miR-339-3p in patients affected with senile cardiac amyloidosis compared to controls. This corresponded to the array profiling results. In contrast, there was no deregulation in the other patient groups. Conclusion MiR-339-3p was increased in blood of patients with senile cardiac amyloidosis. Therefore, miR-339-3p is a potential candidate as biomarker for senile cardiac amyloidosis in future studies. Larger patient cohorts should be investigated.


Introduction
Amyloidosis is caused by dysregulation of protein folding resulting in systemic or organ specific amyloid aggregation. When affecting the heart, amyloidosis can cause severe heart failure, which is associated with a high morbidity and mortality. Different subtypes of cardiac amyloidosis exist e.g. transthyretin cardiac amyloidosis and senile cardiac amyloidosis. Today, diagnostics is primarily based on cardiac biopsies and no clinically used circulating blood-based biomarkers existing. Therefore, our aim was to identify circulating microRNAs in patients with different forms of amyloidosis.

Results
The validation experiments revealed a significant upregulation of miR-339-3p in patients affected with senile cardiac amyloidosis compared to controls. This corresponded to the array profiling results. In contrast, there was no deregulation in the other patient groups.

Introduction
Amyloidosis is a rare and very heterogeneous disease, which can affect several human organs. Pathophysiological amyloidosis is caused by misfolding of non-soluble proteins, which deposit in an extracellular manner. When affecting the heart, cardiac amyloidosis can lead to to cardiac hypertrophy [1,2]. The most common type in the cardiac form is induced by light chain immunoglobulins (AL amyloidosis). In general, the AL type manifests systemically affecting also e.g. the liver or kidney. Another form of this disease is the transthyretin type amyloidosis also called ATTR amyloidosis. A different type affecting mainly men older than 70 years is the wild-type ATTR amyloidosis also named senile cardiac amyloidosis (SCA) [3]. For diagnosis of amyloidosis, cardiac biopsy and echocardiography are used. Until now, there is not any specific biomarker in blood known [3,4]. MicroRNAs (miR's, miRNAs) are small ribonucleic acids about 20 nucleotides. As important regulators they orchestrate post-transcriptionally gene-expression and are therefore involved in cell differentiation processes [5]. In cardiac diseases they are important players in pathologic processes like e.g. cardiac fibrosis and hypertrophy [6]. Therefore, microRNAs can be used as therapeutic targets. More and above, microRNAs circulating in the blood are used as biomarkers for example in Takotsubo syndrome or Hypertrophic cardiomyopathy [7,8]. The aim of our study was to find specific blood-based microRNAs in serum of patients with different forms of cardiac amyloidosis.

Study population
All patients including control and heart failure individuals were recruited at the Columbia University Medical Center in New York. All patients gave written informed consent to participate in the study, which was conducted in accordance with the protocol approved by the Institutional Review Board at Columbia University Medical Center. The diagnosis of different types of amyloidosis was defined by myocardial biopsy and echocardiography. In the heart failure group (HF) only patients with an ejection fraction (EF) < 35% and without amyloidosis or other genetic diseases were included.

RNA isolation from patient blood samples
To obtain liquid supernatant by separating the corpuscular components, the serum samples were centrifuged for 10 minutes at 2000 x g at room temperature. Afterwards the supernatant was stored at -80˚C. The miRNeasy Mini Kit (Qiagen) was utilized according to instructions for the RNA isolation. For subsequent normalization synthetic Caenorhabditis elegans miR-39 (5 μl of 1 fmol/μl) was added as spike-in control during RNA isolation to the Qiazol/chloroform/plasma mixture [8].

TaqMan ArrayCard microRNA screening
For performing a microRNA expression profile isolated RNA of single individuals were pooled within their groups. After preamplification, TaqMan gene expression array cards-384-well microfluidic cards (ThermoFisher SCIENTIFIC) were used according to manufacturer's instructions. 753 different reactions of microRNA were measured. For analyzing array data we used -ddCT method.

Statistical analysis
For statistical analysis and graphical illustration of demographics and expression data Graph Pad Prism 6 was utilized. For analyzing the regulation of miRNAs between the different

Clinical characteristics
In this study, 10 controls, 10 heart failure patients, 13 TTR mutation patients and 11 senile cardiac amyloidosis patients were included. There were no statistical significant differences in age of controls, heart failure and TTR group but the patients affected with senile cardiac amyloidosis were significantly older than control and TTR group. Comparing the gender, in all groups more men than women were included (Table 1). There was no difference in BMI. In the heart failure group we included more patients with diabetes mellitus but this was without statistical significance. Comparing cholesterol, in particular LDL and HDL cholesterol, there was also no difference ( Table 2).

Validation data
Looking at the validated microRNA selection, there was no dysregulation of the circulating miR-99a-5p and miR-122-5p between the different groups (Fig 4A and 4B). Beside it, the circulating expression of miR-27a was not increased in any group, but significantly decreased in TTR patients compared to control (Fig 4C). There were also no differences in circulating levels of the microRNAs miR-221-3p, miR-1180-3p, miR-155-5p and miR-342-3p (Fig 4D, 4E, 4F and 4G). In addition, levels of miR-339-3p were significantly increased in patients with senile cardiac amyloidosis whereas not in the other patient groups (Fig 5). Primer sequences of validated microRNAs are shown in Table 3.

Discussion
In this study, we performed a microRNA assay in samples derived from patients with different forms of cardiac amyloidosis and validated promising biomarker candidates to distinguish different forms of amyloidosis, heart failure patients as well as healthy individuals. Up to date, there is no other study showing the regulation of microRNAs in cardiac amyloidosis. In array profiling experiments, we found 10 candidates being deregulated in TTR amyloidosis and senile cardiac amyloidosis. MicroRNA miR-99a, which was upregulated in array profiling in patients with SCA, was recently shown as tumor suppressor in lung cancer working together with oncogenic proteins EMR2 and E2F2 [9]. In validation experiments, this miRNA failed as possible biomarker for cardiac amyloidosis. Also, miR-574-3p and miR-329 was detected only in a very low number of patient samples. Moreover, miR-122, which was previously described as marker of neurological damage in blood of patients after cardiac arrest [10], did not show any significant regulation in patients with cardiac amyloidosis. Further, miR-155 also failed in our validation cohort as potential biomarker for cardiac amyloidosis. In heart diseases, miR-155 is described as a promoter of cardiac hypertrophy leading to heart failure [11]. Interestingly, we found that miR-339-3p was significantly upregulated in both array data and validation analysis compared to control samples. Until today, miR-339-3p has not been described as a specific circulating biomarker in any heart disease. There is also no previously reported association between miR-339-3p and amyloidosis. It is known that miR-339-3p functions as a suppressor in melanoma by targeting the oncogene MCL1 [12]. Besides melanoma, miR-339-3p is also described as inhibitor for colorectal cancer tumor growth in in-vitro experiments [13]. Thus, miR-339-3p was mainly known in the cancer field. However, our work suggests that miR-339-3p will be a promising candidate as clinical biomarker for senile cardiac amyloidosis. The exact pathophysiological mechanisms of the origin and secretion of miR-339-3p into the blood in cardiac amyloidosis remains unclear and needs to be explored in future studies. Although, miR-221 was described as promoter of cardiac hypertrophy [14], it was not dysregulated in our validation experiments. Also expression levels of miR-27a described e.g. as anti-hypertrophic in rat cardiomyocytes [15] and miR-1180-3p were not significantly dysregulated in any of our patient groups. Interestingly, miR-342-3p being described as decreased in acute heart failure did not show any dysregulation either in our heart failure group or in cardiac amyloidosis groups [16,17].
In this study are also limitations. The unbalanced age of the senile cardiac amyloidosis group is most likely due to the incidence of the disease in higher ages. Another limitation is the small sample size within the different groups owing to the relative rareness of cardiac amyloidosis and therefore the small patient population. In addition, our study did not reveal the pathomechanism behind the upregulation of miR-339-3p in patients with senile cardiac amyloidosis.

Conclusion
In conclusion, miR-339-3p was increased in the plasma of patients with senile cardiac amyloidosis as identified by array profiling and validated by qPCR. Thus, miR-339-3p is a potential candidate biomarker for senile cardiac amyloidosis. Larger cohorts should be investigated in the future.