The occurrence of severe graft-versus-host disease (GVHD) after allogeneic hematopoietic stem cell transplantation (HSCT) leads to a poor prognosis. Recently, with this expansion in the utilization of allogeneic HSCT, such as the introduction of reduced-intensity conditioning and donor lymphocyte infusions, late-onset acute and chronic GVHD has been reclassified based on the exact clinical manifestation instead of the time of onset [1
]. According to the new classification using National Institutes of Health (NIH) consensus criteria established in 2005, patients with de novo acute GVHD occurring beyond day 100 post-transplantation are considered to have late-onset acute GVHD (LA GVHD) [1
]. In accordance with the NIH definition, 20–40% of chronic GVHD cases are reconsidered to be LA GVHD. A few reports have shown that LA GVHD is also associated with greater nonrelapse mortality after allogeneic HSCT than non-GVHD [2
]. However, there is still an insufficient understanding of the association between LA GVHD and a lower incidence of relapse caused by the induction of a beneficial graft-versus-tumor (GVT) effect. To improve long-term survival, intensive efforts have been made to prevent GVHD without losing the GVT effect. Thus, several molecular biomarkers for diagnosis and prediction of the severity and prognosis of GVHD have been explored [4
MicroRNAs (miRNAs) are small noncoding RNA molecules that regulate post-transcriptional gene expression by degrading or suppressing their target mRNA [8
]. Recently, it has been widely recognized that some circulating miRNAs are included in the content of endosome-derived extracellular vesicles including exosomes. Exosomes are secreted by many types of cells into various biological fluids, such as serum and plasma [9
]. The demonstration that exosomes contain not only proteins, RNAs, and lipids, but also functional miRNAs indicated that miRNAs could participate in intercellular communication and elicit immune responses by their transport between cells via exosomes [11
]. Therefore, analysis of the alterations of exosomal miRNAs in the context of various diseases has the potential to facilitate molecular diagnosis [13
In recent studies, miRNA profiling identified a GVHD-specific miRNA signature that can control the regulation of classic acute GVHD in allogeneic immunity. In particular, miR-155, known as an immune-related microRNA, and miR-15a/16 and miR-17-92a were shown to be closely involved in the pathophysiology behind the development of classic acute GVHD in a xenogeneic GVHD murine model [15
]. Moreover, recent studies have demonstrated that specific plasma miRNA signatures, such as miR-423, miR-199a-3p, miR-93, miR-377, and miR-586, could serve as independent biomarkers for the prediction, diagnosis, and prognosis of hematologic diseases and classic acute GVHD [18
]. Despite the fact that the roles of miRNAs in classic acute GVHD after allogeneic HSCT have been extensively studied over the last few years, the relevance and expression profile of miRNAs in LA GVHD have not yet been elucidated. Because a number of previous studies investigated the role of miRNAs as potential predictive biomarkers in acute GVHD, we aimed to determine the clinically relevant exosomal miRNA profile in patients developing LA GVHD after allogeneic HSCT.
Recently, the definition of GVHD was revised in the NIH consensus criteria, with the new category of “late-onset” acute (LA) GVHD being established. Given the short history of this category, the pathophysiology of LA GVHD is still poorly understood. Although the diagnosis of GVHD is judged based on clinical manifestations and pathological findings of the tissue sections from the intestines or liver, it is difficult to provide an accurate pathological diagnosis of GVHD. For this reason, pathological findings have greatly affected various factors of radiation damage, drug injury, and viral infection, presenting a diagnostic dilemma. Therefore, there is an urgent need of exploring reliable, noninvasive biomarkers of GVHD. In the context of GVHD after allogeneic HSCT, biomarker research has progressed through a variety of techniques, such as immunocompetent cells (CD4 helper T cells, regulatory T cells), plasma (soluble IL-2R), and inflammatory cytokines (IFN-γ, TNF-α). For example, Paczesny et al. identified that a panel of four biomarkers (IL-2Rα, TNFR1, IL-8, HGF) in serum was useful for the diagnosis of acute GVHD [4
]. This group also found that the suppression of tumorigenicity 2 (ST2) could distinguish treatment-resistant GVHD patients, and that ST2 levels were associated with nonrelapse mortality [6
]. Moreover, a specific plasma miRNA signature for classic acute GVHD was recently determined by Xiao et al. [18
]. To date, the identification of altered miRNAs has been thought to contribute to the HSCT-related pathogenesis of acute GVHD; however, the mechanism underlying LA GVHD remains to be elucidated. The above-mentioned results provided us with a potential tool as a new liquid biopsy technique.
Because a number of previous studies investigated the role of exosomes and exosomal miRNAs as potential predictive biomarkers in acute GVHD [20
], we aimed to elucidate the possible role of miRNAs specifically in LA GVHD patients. Exosomes containing miRNAs reflect the behavior of certain types of cells, so a survey of changes of miRNAs in exosomes should promote our understanding of clinical conditions. In this context, technical advances in exosome separation and characterization have also recently been achieved, enabling exosomal miRNAs to be easily detected by high-throughput techniques [29
]. Additionally, Kordelas et al. reported an attractive treatment using mesenchymal stem cell (MSC)-derived exosomes for therapy-refractory GVHD [30
]. We also previously reported that many miRNAs have been identified as noninvasive, diagnostic, prognostic, and predictive markers for hematologic malignancy [31
]. In the current study, we explored the LA GVHD-specific exosomal miRNA signature for LA GVHD compared with that in non-GVHD. We also performed bioinformatic analysis to investigate the potential molecular mechanisms behind LA GVHD and identify molecular target genes. Our results of validation by qRT-PCR showed that miRNA-128 significantly increased in LA GVHD patients. The miRNA miR-128 is well known as a brain-enriched miRNA, which plays a crucial role in the development of the nervous system [34
]. The miRNA miR-128 has also been previously demonstrated to be dysregulated in tissues and blood samples in several malignant tumor patients and to be correlated with tumor progression [35
]. In the current study, it was impossible to conduct a functional analysis of miR-128. Hence, using MirTarBase, we instead searched for potential target genes to investigate the mechanism of miR-128 (Table 3
We perfomed ROC analysis to assess the diagnostic capacity of exosomal miRNAs. Exosomal miR-128 exhibited its potential in discriminating LA GVHD from non-GVHD groups. When the expression level of candidate miRNAs was analyzed sequentially in a subset of LA GVHD patients during allogeneic HSCT, miR-128 was shown to increase over the cut-off value earlier than the development of the disease. However, our results must be interpreted with caution, because this study was limited by the relatively small number of samples from LA GVHD patients and absence of a validation cohort of any sort. After all, molecular mechanisms on the dysregulation of exosomal miRNAs in patients with LA GVHD could not be fully provided. Also, candidate miRNAs were then examined longitudinally in only two patients over time. Thus, these results need to be further studied with large cohort.
The results of the top 10 miRNAs obtained by TLDA included several characterized miRNAs (miR-423-5p, miR-19a, and miR-142) known to be dysregulated in acute GVHD [18
]. In particular, miR-423-5p, which is involved in the immune response of acute GVHD incidence, was elevated up to 49-fold in exosomes of LA GVHD patients in our study. Crossland et al. reported differential expression of miR-423 in exosomes at day 14, after HSCT had shown potential as a predictive biomarker for the occurrence of acute GVHD [20
]. Therefore, the pathology of LA GVHD may resemble that of acute GVHD caused by donor T-cell alloreactivity in light of miRNAs. However, the biology of GVHD is complex. In other words, miRNAs also regulate various molecular targets, including normal development, differentiation, and maturation of hematopoietic cells in the immune system [38
]. The difference of miRNA expression may influence the process of immune reconstitution after allogeneic HSCT. Moreover, the changes of immune-related miRNAs may be due to the dosage of immunosuppression. The dysregulated miRNAs when LA GVHD occurs may be repressed by immunosuppressive regimen. However, it is difficult to evaluate whether the level of this miRNA returns to baseline upon therapeutic intervention because of the uncontrollable nature of LA GVHD. It is a challenge for future study to clarify the association between miRNA expression and response to therapy. Although it is difficult to define miRNAs specific to LA GVHD in very complex medical cases, and in the small number of transplanted patients in the current study, exosomal miR-128 may be related to the immunoregulation of LA GVHD.
In summary, this study demonstrated that the levels of several exosomal miRNAs change at the time of LA GVHD onset after HSCT. Although a longitudinal series of blood samples was taken from 2 patients only, upregulation of miR-128 expression level preceded of LA GVHD incidence. In particular, our results suggest that biological alterations of miR-128 in exosomes may serve as potent predictive biomarkers for close monitoring of the onset of LA GVHD. We consider that this work could provide a foundation for revealing the importance of miRNA in the pathophysiology of LA GVHD. GVHD remains a major cause of transplant-associated complications, which can lead not only to lethal organ damage, but also to a decreased quality of life. Inhibition of certain exosomal miRNAs, such as miR-128, might be considered as a novel potential treatment of LA GVHD to target the immune system and inflammatory signaling, to control GVHD. However, further studies are needed to elucidate the molecular mechanisms behind the development of LA GVHD.
4. Materials and Methods
The use of patient samples was approved by the Institutional Review Board of Tokyo Medical University (IRB no. 1979, approved on 28 March 2011). Written informed consent was obtained from all of the participants before the collection of specimens, in accordance with the Declaration of Helsinki.
4.2. Diagnosis of Late-Onset Acute GVHD
The diagnosis of LA GVHD here was based on clinical symptoms or was histologically proven by biopsy in the target organs, as described previously [1
]. Patients with typical manifestations of classic acute GVHD occurring beyond 100 days after transplantation and persistent, recurrent acute GVHD were considered as having LA GVHD, in accordance with the new classification based on NIH consensus criteria [1
4.3. Isolation of Exosome Fractions from Plasma Sample of Patients
Ten patients with hematological malignancies who underwent allogeneic HSCT between February 2012 and November 2013 were included in this study, which consisted of five patients with LA GVHD (gut, n
= 2; liver, n
= 2; skin and gut, n
= 1) and five stable patients without GVHD symptoms (non-GVHD) under immunosuppression. All samples were collected sequentially after HSCT. For patients with LA GVHD, we evaluated blood samples obtained at the onset of this disease. No treatment for LA GVHD had been given at the time of sampling. For comparison, blood samples from HSCT patients without GVHD symptoms and other complications were used. In addition, eight healthy controls were also randomly selected in our hospital. We extracted exosome fractions from 200 µL of plasma using Total Exosome Isolation Reagent (Invitrogen, Carlsbad, CA, USA) and then assessed exosomal miRNA expression in a subset of samples, as described previously [11
4.4. Exosomal miRNA Profiling
Total RNA was isolated from exosomes fractions of five patients with LA GVHD and five patients without GVHD (non-GVHD) using the miRNeasy Mini kit (Qiagen, Hilden, Germany). Exosomes were diluted with 700 μL of QiaZol. After 5 min of incubation, 1 μL of 1 nM ath-miR-159 (Hokkaido System Science, Hokkaido, Japan) was added to the aliquot followed by vortexing for 30 s and incubation for 5 min. Subsequent phenol extraction and centrifuge filtration were performed, in accordance with the manufacturer’s instructions (Qiagen, Hilden, Germany). The RT reaction and pre-amplification step were set up in accordance with the manufacturer’s instructions (Thermo Fisher Science, Waltham, MA, USA). miRNAs were reverse-transcribed with Megaplex Prime Pools (Human Pools A v2:1; Thermo Fisher Sciences, Waltham, MA, USA).
4.5. TaqMan Loq-Density Array Screening
The miRNA expression profiles were determined with a TaqMan miRNA Array Human Card A (Thermo Fisher Sciences, Waltham, MA, USA). Quantitative RT-PCR was performed on an Applied Biosystem 7900 HT thermocycler, in accordance with the manufacturer’s recommended program [31
]. Using SDS2.2 software and Data Assist (version 3.01, Thermo Fisher Sciences, Waltham, MA, USA), the expression of exosomal miRNAs was calculated based on cycle threshold (Ct) values normalized by those of ath-miR-159, which was spiked in each exosomal sample. Data analysis was performed using GeneSpring®
software (Version 12.1, Agilent Technologies, Palo Alto, CA, USA) and R software (https://www.r-project.org
, accessed on 27 Arpil 2017). The Benjamini–Hochberg algorithm was used for the estimation of false discovery rates, as we reported previously [31
4.6. Real-Time Quantitative RT-PCR for Candidate miRNAs
We determined the levels of the individual miRNAs by real-time quantitative RT-PCR with a TaqMan MicroRNA Assay (Thermo Fisher Sciences, Waltham, MA, USA) and the following miRNA-specific stem-loop primers: hsa-miR-128 (assay ID: 002216, Ambion, Austin, TX, USA) and hsa-miR-125b (assay ID: 000449, Ambion, Austin, TX, USA). Subsequently, quantitative real-time PCR was performed with an ABI Prism 7900 sequence detection system (Thermo Fisher Sciences, Waltham, MA, USA). The reaction was initiated by incubation at 95 °C for 2 min, followed by 50 cycles of 95 °C for 15 s and then 60 °C for 1 min. All reactions were run in duplicate. Mean (Ct) values for all miRNAs were quantified with sequence detection system software (SDS version 1.02; Thermo Fisher Sciences, Waltham, MA, USA) [31
]. The expression of all miRNAs was normalized to miR-159, which was stably detected in all samples, yielding a −ΔΔCt value that was calculated by subtracting the −ΔΔCt value of the normal samples from the respective −ΔCt values of patient samples. The expression of all miRNAs was normalized using the 2−ΔΔCt
Following the identification of differentially expressed miRNAs, the predicted target genes of these altered miRNAs were investigated using the experimentally validated miRNA–target interactions database MirTarBase (http://mrtarbase.mbc.nctu.edu.tw/
, accessed on 11 September 2017).
4.7. Statistical Analysis
Data are expressed as mean ± standard deviation (SD). Multiple groups were compared by one-way analysis of variance (ANOVA) with post hoc Tukey’s test. Statistical analysis was performed using GeneSpring software, and miRNAs with both a ΔCt of either >1.0 or <−1.0 and a P value of less than 0.05 were deemed to be differentially expressed. The area under the curve (AUC) of the receiver-operating characteristic (ROC) curve was calculated. GraphPad Prism software (version 5c for Macintosh; GraphPad Software Inc., La Jolla, CA, USA) was used for statistical analysis.