Alpha-2-HS-glycoprotein plasma level decrease correlates with age in patients with myelodysplastic syndromes

Abstract

BACKGOUND:

It has been indicated in plasma proteomic studies on different myelodysplastic syndrome (MDS) cohorts that alpha-2-HS-glycoprotein could be a promising MDS biomarker candidate.

OBJECTIVE:

The goal of this work was to estimate alpha-2-HS-glycoprotein (AHSG) plasma levels and its biomarker value in the low- and high-risk subgroups of MDS patients.

METHODS:

The level of AHSG was estimated for 115 plasma samples using ELISA.

RESULTS:

The AHSG plasma level was found to be decreased significantly ( $p=$ 2.59 $\times$ 10 ${}^{-7}$ ) in MDS patients (515 $\pm$ 58 $\mu$ g/ml) when compared to healthy controls (579 $\pm$ 64 $\mu$ g/ml). Pearson and Spearman correlation analyses showed that age is the principal factor affecting the AHSG plasma level, rather than risk/diagnosis in MDS.

CONCLUSIONS:

In this work we demonstrate that although the total plasma level of AHSG is decreased in myelodysplastic syndrome patients, in particular in advanced MDS, that decrease correlates more strongly with age than with diagnosis within our studied cohort. Thus, according to the AHSG data gathered so far, AHSG total plasma level does not seem to be a suitable MDS biomarker, but its particular proteoforms should be considered for the next steps in MDS research.

1. Introduction

In recent years we have performed plasma proteomic studies on different myelodysplastic syndrome (MDS) cohorts and found indications that alpha-2-HS-glycoprotein (AHSG; UniProt P02765) could be a promising MDS biomarker candidate. Our results, obtained by 2D SDS-PAGE together with ELISA plasma level estimations for refractory anemia (RA) [1] and RAEB-2 [2] patient cohorts, showed AHSG plasma level decreases for both MDS cohorts compared to healthy individuals. In addition, mass spectrometry-based relative quantification indicated the presence of posttranslational modification(s) (PTMs). The potential for biomarker utilization was also supported by the previous observation that AHSG regulated tumor growth in mice, and thus may be involved in the pathophysiology of malignant diseases [3]. AHSG antibodies were found in the serum of breast cancer patients [4], and AHSG was proposed as a serum marker of survival in patients with glioblastoma, with its serum level decrease correlating with the degree of tumor malignancy [5]. Moreover, in our recent re-analysis of the proteomic data of different MDS cohorts covering low- to high-risk subgroups, AHSG was identified as a possible biomarker [6]. Using a technique (western blotting) additional to that previously employed, we have again shown the AHSG plasma decrease in advanced MDS subgroups, confirming the previous results. In this work our goal was to estimate AHSG plasma levels and its biomarker value in the low- to high-risk subgroups of MDS using larger patient cohorts.

Table 1

Characteristics of patient and control cohorts with AHSG plasma level results

	$n$	age – avg	age – range	sex (m/f)	AHSG ( $\mu$ g/mL) – avg	AHSG ( $\mu$ g/mL) – std
Cy	11	25.4	20–36	4/7	610	80
Co	34	56.1	50–65	15/19	568	55
C	45	48.6	20–65	19/26	579	64
MDS low-risk	55	57.8	21–87	26/29	524	56
MDS high-risk	15	64.1	49–79	6/9	485	56
MDS	70	59.2	21–87	32/38	515	58

Cy – younger controls; Co – older controls; C – controls; AHSG – alpha-2-HS-glycoprotein.

2. Methods

The level of AHSG was estimated with an ELISA kit (ab108855; Abcam) according to the manufacturer’s instructions. In total, 115 plasma samples were analyzed: 45 healthy control (11 younger and 34 older donors) and 70 MDS patient (55 low-risk and 15 high-risk MDS) samples. The healthy control samples were divided into younger (as used in the original proteomic studies) and older (of comparable age to the patients with MDS) donors in order to follow the previous study design for retinol binding protein 4 [7]. Characteristics of patient and control cohorts together with AHSG plasma level results are summarized in Table 1. All samples were obtained and analyzed in accordance with the Ethical Committee regulations of the Institute of Hematology and Blood Transfusion in Prague and with the Helsinki Declaration. Informed consent was obtained from all individual participants included in the study. A t-test with the Bonferroni corrected p-value threshold ( $p_{Th}=$ 0.005; $\alpha =$ 0.05) was used to compare the AHSG plasma level between different cohorts. In order to establish the influences of different factors (age, sex, and risk/diagnosis – low-risk and high-risk cohorts), Pearson and Spearman correlation analyses were performed; the 0.01 level (2-tailed) was a threshold for a correlation to be significant. General linear model and non-parametric correlation coefficient were used. Univariate Analysis of Variance was further performed to differentiate the influence of each factor. IBM SPSS software and R sortware (http://www.R-project.org) [8] were used for statistical analyses.

3. Results

The AHSG plasma level was found to be decreased significantly ( $p=$ 2.59 $\times$ 10 ${}^{-7}$ ) in MDS patients (515 $\pm$ 58 $\mu$ g/ml) when compared to healthy controls (579 $\pm$ 64 $\mu$ g/ml). More precisely, the plasma levels were 610 $\pm$ 80 $\mu$ g/ml and 568 $\pm$ 55 $\mu$ g/ml for younger and older controls, respectively. The results of AHSG plasma level measurement in the studied cohorts are summarized in Table 1. A non-significant decrease was observed in plasma level in the high-risk MDS, versus low-risk MDS patients (485 $\pm$ 56 $\mu$ g/ml and 524 $\pm$ 56 $\mu$ g/ml, respectively; $p=$ 0.02). For MDS patients over 60 years of age there was an AHSG plasma level decrease (491 $\pm$ 48 $\mu$ g/ml) compared to MDS patients below 60 years of age (545 $\pm$ 55 $\mu$ g/ml); $p=$ 3.79 $\times$ 10 ${}^{-5}$ . In order to establish the influences of age and risk/diagnosis factors (low-risk and high-risk cohorts), Pearson and Spearman correlation analyses were consequently performed. Both analyses showed that age significantly influenced AHSG plasma levels in the studied cohorts: Pearson’s $r=$ 0.417 ( $p=$ 0.003) and Spearman’s $\rho =$ 0.456 ( $p<$ 0.001). There was no significant influence ( $p<$ 0.01) of other studied factors observed; risk factor did not reach the significancy threshold but satisfied $p<$ 0.05 in both correlation analyses. Consequently, Univariate Analysis of Variance was performed and age was confirmed as the influencing factor (for age, sex, and risk was $p=$ 0.005, $p=$ 0.828, and $p=$ 0.058, respectively).

4. Discussion

Our previous studies [1, 2, 6] showed that AHSG (based on SDS-PAGE and mass spectrometry data) is altered in MDS patients and seemed to be a promising MDS marker. Nevertheless, the methods used were not able to unambiguously answer whether the cause was a change in the total AHSG plasma level or in its posttranslational modification(s). Some clues suggested AHSG modifications (western blot, mass spectrometry-based relative label-free quantification of tryptic peptides) while ELISA the plasma level change. Therefore, we decided to determine the influence of the plasma level using enlarged patient and control cohorts; in this work, the total AHSG plasma level was found to be decreased in MDS patients (compared to healthy controls), especially in the high-risk MDS cohort. However, since the high-risk MDS patient cohort was of higher average age than the low-risk patients, and an AHSG plasma level decrease was observed in older healthy controls compared to younger ones, we estimated whether or not age could be a factor influencing AHSG plasma levels. Three different factors (age, sex, and risk/diagnosis) were estimated; only age was observed to significantly influence the total AHSG level. Therefore, age is the principal factor affecting the AHSG plasma level, rather than risk/diagnosis in MDS (within the studied cohorts).

To complement our interpretation of the results, limitations of this work should be understood. First, AHSG has been described to be involved in other diseases and complications that are in particular associated with higher age (e.g., diabetes [9], insulin resistance [10], atherosclerosis development [11]). Therefore, comorbidities related to higher age, instead of age itself, could be the crucial controlling factor. Although the control and patient cohorts were enlarged in this work compared to our previous report on retinol binding protein 4 plasma level alterations [7], it is still far too small to be suitable for proper statistical analysis including comorbidity-related parameters. Second, in our previous work [6] it had been shown using western blotting that AHSG is present in the plasma of MDS patients in several forms of different molecular weights, and that some of these AHSG forms (proteoforms) differ more substantially among the MDS subgroups than the total AHSG level itself. The most evident difference was observed for the RAEB-1 subgroup, interestingly the only subgroup in which AHSG was not identified as one of the alternating proteins when comparing this subgroup with the healthy controls [12]. Therefore, more effort should be focused on the identification and characterization of all AHSG proteoforms and their profiling in MDS. According to the data on AHSG that have now been gathered, AHSG total plasma level does not seem to be a suitable MDS biomarker; its particular proteoforms instead of the total plasma level should be considered for the next steps in MDS research.