Abstract
B-cells play a significant role in chronic graft-versus-host disease (cGVHD). Male patients with female donors (F→M) are at a higher risk of developing cGVHD. B cell responses against minor histocompatibility antigens encoded on the Y chromosome, called H-Y antigens, develop following F→M HCT patients in association with cGVHD (Miklos, Blood. 2005 & Sahaf, PNAS. 2013). Here we present our novel HY microarray and use this sensitive technology to determine temporal development of HY antibody (Ab) preceding cGVHD. Multivariate analyses demonstrate that HY-Ab detection 3 months (3m) post HCT predicts cGHVD incidence and non-relapse mortality (NRM).
We studied 136 adult male recipients of F→M HCT between 2005 and 2012 who survived without relapse for at least 3m post-HCT with 3m plasma available. Median patient age was 53 (21-74). Related donors were transplanted in 85 (63%) and 128 (94%) were PBSC grafts. Reduced intensity conditioning accounted for 61 (45%) and anti-thymocyte globulin (ATG) was used in 71 (52%). Thirty-one patients (23%) experienced grade II-IV acute GVHD.
We measured IgG against six HY antigens (DBY, UTY, ZFY, SMCY, EIF1AY, and RBS4Y) from plasma collected 3m post-HCT using a novel proteomic microarray here presented for the first time. The cut-off value for seropositivity was defined as the third quartile + 2x the interquartile range, determined from plasma of 60 male donors. HY-score was defined as the cumulative number of HY antigen targeted by Abs at 3m post-HCT.
The frequencies of HY antigen-specific Ab are presented in Table 1, showing that SMCY and UTY were most frequently detected and overall, 78 (57%) had developed allo-Ab against any of these 6 HY antigens. Each HY-Ab was significantly associated with the development of cGVHD and DBY was greatest. LASSO analysis suggested that DBY, UTY, and ZFY were the most predictive for the development of cGVHD (Table 1). Univariate analysis failed to identify associations between clinical features and the development of HY-Ab at 3m. The detection of HY-Ab gradually increased within the 1st year post HCT and seropositivity for each HY-IgG (except RPS4Y) persisted.
Considering each HY-IgG response by principal component analysis, a higher HY-score was associated with an increased risk for the development of cGVHD and NRM, after adjusting for usual alloHCT clinical factors (Table 2). In addition, the severity of cGVHD was significantly associated with the HY-score: the proportion of severe/moderate cGVHD was 33% in 0, 30% in 1, 60% in 2-3, and 70% in 4-6 (P<0.01).
. | DBY . | UTY . | ZFY . | SMCY . | EIF1AY . | RPS4Y . | Any-HY . |
---|---|---|---|---|---|---|---|
At 3m post-HCT (n=136) | 26 (19%) | 45 (33%) | 21 (15%) | 52 (38%) | 27 (20%) | 25 (18%) | 78 (57%) |
>1y post-HC T(n=91) | 36 (40%) | 39 (43%) | 24 (26%) | 44 (48%) | 26 (29%) | 38 (42%) | 65 (71%) |
Association between HY-IgGat 3m and >1y | P=0.05 | P<0.01 | P=0.02 | P<0.01 | P<0.01 | P=0.15 | - |
CGVHD incidence at 1y | |||||||
3m positive | 69% | 62% | 71% | 64% | 63% | 62% | 56% |
3m negative | 42% | 40% | 43% | 37% | 43% | 44% | 34% |
P-value | P<0.001 | P<0.01 | P<0.001 | P<0.01 | P<0.01 | P=0.03 | P=0.03 |
LASSO coefficient | 0.32 | 0.2 | 0.25 | 0.1 | 0.1 | 0 | - |
. | DBY . | UTY . | ZFY . | SMCY . | EIF1AY . | RPS4Y . | Any-HY . |
---|---|---|---|---|---|---|---|
At 3m post-HCT (n=136) | 26 (19%) | 45 (33%) | 21 (15%) | 52 (38%) | 27 (20%) | 25 (18%) | 78 (57%) |
>1y post-HC T(n=91) | 36 (40%) | 39 (43%) | 24 (26%) | 44 (48%) | 26 (29%) | 38 (42%) | 65 (71%) |
Association between HY-IgGat 3m and >1y | P=0.05 | P<0.01 | P=0.02 | P<0.01 | P<0.01 | P=0.15 | - |
CGVHD incidence at 1y | |||||||
3m positive | 69% | 62% | 71% | 64% | 63% | 62% | 56% |
3m negative | 42% | 40% | 43% | 37% | 43% | 44% | 34% |
P-value | P<0.001 | P<0.01 | P<0.001 | P<0.01 | P<0.01 | P=0.03 | P=0.03 |
LASSO coefficient | 0.32 | 0.2 | 0.25 | 0.1 | 0.1 | 0 | - |
. | HY-score . | n . | Incidence . | P-value . | HR (95%CI) . | P-value . |
---|---|---|---|---|---|---|
CGVHD at 1y | 0 | 58 | 34% | <0.001 | 1 | ref |
1 | 30 | 37% | 0.90 (0.43-1.91) | 0.79 | ||
2-3 | 25 | 65% | 2.08 (1.13-3.85) | 0.019 | ||
4-6 | 23 | 74% | 4.71 (2.23-9.91) | <0.01 | ||
NRM at 3y | 0 | 58 | 1.8% | 0.04 | 1 | ref |
1 | 30 | 8.9% | 0.93 (0.09-9.53) | 0.95 | ||
2-3 | 25 | 16% | 3.77 (0.77-18.4) | 0.1 | ||
4-6 | 23 | 29% | 10.6 (1.28-87.2) | 0.029 |
. | HY-score . | n . | Incidence . | P-value . | HR (95%CI) . | P-value . |
---|---|---|---|---|---|---|
CGVHD at 1y | 0 | 58 | 34% | <0.001 | 1 | ref |
1 | 30 | 37% | 0.90 (0.43-1.91) | 0.79 | ||
2-3 | 25 | 65% | 2.08 (1.13-3.85) | 0.019 | ||
4-6 | 23 | 74% | 4.71 (2.23-9.91) | <0.01 | ||
NRM at 3y | 0 | 58 | 1.8% | 0.04 | 1 | ref |
1 | 30 | 8.9% | 0.93 (0.09-9.53) | 0.95 | ||
2-3 | 25 | 16% | 3.77 (0.77-18.4) | 0.1 | ||
4-6 | 23 | 29% | 10.6 (1.28-87.2) | 0.029 |
*HR was obtained by multivariate analysis after adjusting for age of pts, disease risk, CMV status, donor type, mobilized graft, ATG, TBI, and grade of aGVHD
Receiver operating characteristic (ROC) curve analysis revealed that HY-score in combination with clinical factors enhanced the predictive potential for the development of cGVHD [area under the curve (AUC): 0.76], in comparison with either of only HY-score (AUC: 0.66) or clinical factors (AUC: 0.69).
Here, we show that HY Ab detection 3m following sex-mismatch HCT actually predicts the development of cGVHD, independently from clinical risk factors. In addition, the combination of HY-score and clinical factors had a greater predictive potential than clinical factors alone for the development of cGVHD in F→M HCT. HY-Ab development 3m post HCT may stratify cGVHD risk and support B-cell-depletion therapy beginning 3 months or earlier to prevent cGVHD development.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
This icon denotes a clinically relevant abstract
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal