RNA sequencing suggests that non‐coding RNAs play a role in the development of acquired haemophilia

Abstract Acquired haemophilia (AH) is a rare disorder characterized by bleeding in patients with no personal or family history of coagulation/clotting‐related diseases. This disease occurs when the immune system, by mistake, generates autoantibodies that target FVIII, causing bleeding. Small RNAs from plasma collected from AH patients (n = 2), mild classical haemophilia (n = 3), severe classical haemophilia (n = 3) and healthy donors (n = 2), for sequencing by Illumina, NextSeq500. Based on bioinformatic analysis, AH patients were compared to all experimental groups and a significant number of altered transcripts were identified with one transcript being modified compared to all groups at fold change level. The Venn diagram shows that haemoglobin subunit alpha 1 was highlighted to be the common upregulated transcript in AH compared to classical haemophilia and healthy patients. Non‐coding RNAs might play a role in AH pathogenesis; however, due to the rarity of HA, the current study needs to be translated on a larger number of AH samples and classical haemophilia samples to generate more solid data that can confirm our findings.


| INTRODUC TI ON
Acquired haemophilia (AH) is a very rare disease, in which autoantibodies produced in an associated underlying condition inactivate FVIII. 1 Cutaneous purpura and internal bleeding are the main clinical manifestations. 2 Diagnosis is based on measuring FVIII concentration in plasma, in the presence of inhibitors, while the treatment is mostly focused on inhibiting the bleeding episodes. 3,4 In AH, epigenetics could play an important role in both deciphering non-mutational and mutational status of F8 gene, as well as how autoimmunity is triggered against FVIII. As noncoding molecules can interact with F8 gene and non-genetic events may cause AH, non-coding RNAs potentially influence F8 expression and can trigger autoimmunity against FVIII. However, the impact of non-coding RNAs in F8 gene modulation is still unclear, due to the contradictory results previously reported. 5,6 The methylation pattern of F8 gene and the surrounding regions could be attributed to HA severity.
Yes, studies on family members with F8 mutations showed that they express different degrees of severity, while other studies showed no difference between controls and haemophilic patients regarding their methylation pattern. Thus, due to the high variability in F8 mutational status, the link between epigenetics and HA must be further investigated and certain interactions must be confirmed. [7][8][9][10] In our study, we investigated the transcriptomic signature in RNAs from the plasma of two AH patients, six classical haemophilia patients (three mild and three severe) and compared them to two healthy donors; by RNA-sequencing to highlight relevant differentially expressed transcripts. We observed that haemoglobin subunit alpha 1 and both coding and non-coding RNAs show different expressions in AH compared to all other investigated groups, potentially being a reliable biomarker candidate for predicting AH development.
This might provide an insight into the altered biological pathways involved in AH development, which may be subject for further studies.

| Sample collection
As previously published protocols for serum and plasma isolation, the peripheral blood was collected on K 2 EDTA-coated vacutainer and the plasma was immediately separated by and stored at −80°C. 11,12 Both AHA cases had no pre-existent comorbidities and were diagnosed as idiopathic AHA, not related to any underlying condition.

| RNA sequencing
Small RNA Seq libraries were generated using the 'TrueQuant smallRNA-Seq Kit' for ultra-low input material by GenXPro GmbH in Frankfurt am Main, Germany, according to the manual of the manufacturers. After silica column-based isolation of small RNA from 500 μL of EDTA-blood-plasma, adapters including unique molecular Identifiers (patent number 102008025656) were ligated to both ends of the RNA. After cDNA generation and second-strand synthesis, PCR with minimum number of cycles was used to produce a library that was sequenced on an Illumina NextSeq500 machine with 1× 75 bps.

| Bioinformatics workflow for small-RNA seq
Raw data was pre-processed using Cutadapt 13 to eliminate lowquality reads. FastQC was used to assess the quality of sequencing after trimming. Cleaned reads were mapped to the human reference genome GRCh38 using Bowtie2 14 and to the miRbase (https:// www.mirba se.org/). Quantification of mapped reads to each gene was performed using HT-seq. 15 Differential expression analysis was performed using DESeq2, 16 which is based on negative binomial generalized linear models. Results were compiled into a final

| Statistical analysis
The normalized values were compared using Welch's t-test. A comparison was considered significant if the log2 FC was over one and the unadjusted p-value was under 0.01.

| RE SULTS
The transcriptomic signature in the RNA isolated from plasma of AH patients (n = 2; Aquired_1, Acquired_2), and mild classic haemophilia  Figure 3A.
A significant number of altered transcripts were observed when comparing AH to healthy donors and classical haemophilia patients.
As presented in Figure 3A, the cutoff of the FC > 2 was used to select the most relevant differentially expressed transcripts. Based on the different transcript patterns it is visible that AH show different expressions than all other groups, which may lead us to a new path which can indicate potential transcripts that can act as future biomarkers.
After comparing all samples and determining which transcripts are altered between the groups, different subsets of transcripts (both coding and non-coding) may act as start points for future investigations in the associated mechanisms. As represented in Figure 3B, haemoglobin subunit alpha 1 (HBA1-201) is one of the specific upregulated transcripts in AH when compared to all other conditions. HBA1 may indicate a compensatory physiological response to acute blood loss.
HBA1 is located on chromosome 16 and has around 30 kb with two identical coding sequences alpha 2 and alpha 1, encoding the two alpha and betta chains of HbA. Previous data shows that deletions in HBA1 and HBA2 lead to deficit in haemoglobin and thalassemia. 17,18 Each non-coding RNA can interact with multiple genes and one gene can be targeted by multiple other RNAs. In the last years the miRNA mediated mRNA expression was investigated and shown to be a key biological mechanism that is involved in gene expression, as miR-NAs tend to bind o the 3′ untranslated region (3'UTR) of mRNAs and influence gene expression. 19,20 Thus, the interaction between non-coding RNAs and F8 gene might represent an important path in understanding how patients without F8 mutations are lacking FVIII in blood and develop HA. It was shown that miR-374b-5p and miR-30c-5b target F8 gene and impair FVIII, indicating that classical HA could be evaluate not only by the mutational status. 5,21

| DISCUSS ION
In the case of HA, miRNAs shown direct involvement in the downregulation of F8 gene and many studies were indicating that other F I G U R E 1 Principal component analysis of the included samples from AH patients (Aquired_1, Acquired_2), and mild classic haemophilia (Mild_6, Mild_7, Mild_8), severe classic haemophilia (Severe_3, Severe_4, Severe_5) and healthy donors (Healthy_9 and Healthy_10).

F I G U R E 2
Heatmap of the small RNAs that were significantly different between any of the conditions. Comparison between samples from AH patients (Aquired_1, Acquired_2), and mild classic haemophilia (Mild_6, Mild_7, Mild_8), severe classic haemophilia (Severe_3, Severe_4, Severe_5) and healthy donors (Healthy_9 and Healthy_10).
genes that encode coagulation factors can be modulated by miRNAs thus triggering bleeding disorders and thrombosis. 19,22 AH is characterized by the autoantibodies that neutralize FVIII by As presented in Figure 1, the plotting was not significant on PCA due to the low number of samples which represent our main study limitation, however by using the bioinformatic tools we were able to evaluate the most relevant small RNAs identified in each sample. As represented in Figure 2

| CON CLUS ION
While classical haemophilia is X-linked disease and occurs due to mutations in F8 gene, acquired haemophilia is developed due to biological mechanisms that can be epigenetically modulated by different factors such as non-coding small RNAs. Both cases of acquired haemophilia A were idiopathic, with not underlying medical condition. Even if the demographic characteristics of both patients were similar, we completely understand that two cases are not truly relevant to draw any definitive conclusion, but acquired haemophilia A is an ultra-rare disease, especially idiopathic cases, that did not have any pre-existent malignancy, any endocrinologic, infectious or surgical history. Still, investigations regarding this aspect are needed to discover new biomarkers for acquired haemophilia.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data in available upon request.