Discovery and verification of blood-based protein biomarker candidates for prediction of acute graft-versus-host disease and non-relapse mortality by using mass spectrometry-based proteomics approaches
질량분석기 기반의 단백질체학 방법론을 이용하여 이식편대숙주병의 발병위험 및 무재발 사망률 예측을 위한 혈액 단백질 생체표지자의 발굴 및 검증 연구
- 의과대학 임상의과학과
- Issue Date
- 서울대학교 대학원
- 학위논문 (박사)-- 서울대학교 대학원 : 의과대학 임상의과학과, 2018. 8. 김인호.
- Graft-versus-host disease (GVHD) is a major complication of allogeneic hematopoietic stem cell transplantation (allo-HSCT), which is a treatment for different types of malignant and non-malignant hematological disorders. Clinically, acute GVHD (aGVHD) represents a critical barrier to widespread utilization of alloHSCT as a first-line therapeutic option despite its unique curative potential because it is a major cause of non-relapse mortality (NRM) in patients undergoing alloHSCT. Biomarkers for predicting the risk of aGVHD can be used to identify, before the onset of clinical manifestation, high-risk patients who may benefit from early risk-reducing interventions such as preemptive immunosuppressive therapy. Efforts have also been made to identify markers for predicting aGVHD development before the onset of clinical symptoms. However, no single biomarker or composite panel has been established to clearly discriminate between patients who will develop aGVHD and those who will not.
In this study, I investigated potential biomarkers for predicting the risk of aGVHD and NRM using a label-free quantitative mass spectrometry-based proteomic method to identify candidate proteins as biomarkers and then verified the candidates by multiple reaction monitoring (MRM) mass spectrometry using plasma samples collected from patients who underwent allHSCT. In the discovery phase, I compared the proteome profile of pooled plasma obtained from 5 aGVHD-positive patients and 5 aGVHD-negative patients. A total of 202 unique proteins was identified in the two groups. Among them, 16 differentially expressed proteins (DEPs) predicted to be associated with aGVHD development were extracted and subjected to MRM MS-based relative quantification. Thirty-four heavy peptides were used for MRM method development, and the established liquid chromatography (LC)-MRM method was applied to measure the relative protein levels in individual patient samples (n = 10) used in the discovery experiments. Seven candidate proteins with significantly higher levels in the GVHD-positive patient group (beta-2—microglobulin, leucine-rich alpha-2-glycoprotein, epidermal growth factor-containing fibulin-like extracellular matrix protein 1, peroxiredoxin-2, metalloproteinase inhibitor 1, plastin-2, and REG 3α) were subjected to MRM MS-based absolute quantification for verification of the method in an independent patient cohort. A total of 88 multiplexed MRM transitions were established and applied to precisely measure the absolute concentration of 14 candidate peptides in the plasma of 89 patients.
The predictive value of the candidate biomarkers was evaluated in terms of the risk of aGVHD and NRM by constructing an optimal multivariable Cox model containing clinical characteristics and biomarker candidate levels as variables. Patients with high levels of each candidate biomarker showed a consistent tendency towards a higher risk of aGVHD and NRM, as compared to patients with low levels of these markers in the post-engraftment plasma samples of the verification set. TIMP-1, plastin-2, and REG3α were selected and used together to develop a biomarker panel score that ranged from 0 to 3. The biomarker panel score was significantly correlated with the risk of aGVHD and NRM in the univariable and multivariable Cox models. Model performance evaluation based on likelihood ratio test, five-fold cross-validated C (5-CVC) indices, and a continuous form of the survival-based net reclassification improvement (NRI) index demonstrated that addition of the biomarker panel score to clinical predictors significantly improved the discriminatory performance of the Cox model for predicting aGVHD risk and NRM. These findings suggest that plasma-based protein biomarkers can be used to predict aGVHD occurrence and NRM before the clinical onset of symptoms.
- Files in This Item: