High expression of dedicator of cytokinesis 1 adversely influences the prognosis of acute myeloid leukemia patients undergoing allogeneic hematopoietic stem cell transplantation

Background

Acute myeloid leukemia (AML) is a malignant disease of hemopoietic stem cells. It is characterized by clonal expansion of differentiated blasts of myeloid lineage.¹ The current version of the WHO classification provides guidance on AML diagnosis, treatment and prognostication.² The outcome of AML patients is heterogeneous, depending on both patient- and disease-related risk factors. Genomic abnormalities account for at least 60% of the variables in AML prognostication.³ For example, FLT3-ITD has a clear correlation with a poor outcome, while mutations in NPM1 bode a particularly better prognosis.⁴ The presence of biallelic mutations of CEBPA is an independent factor for favorable outcome in AML patients.⁵ A recent meta-analysis indicated that IDH1 mutations confer a poorer survival,⁶ so do genetic aberrations of TP53⁷ or RUNX1.⁸ Genetic factors are incorporated to treatment design because they determine disease sensitivity to treatment as well as the tolerability of therapy.

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) has a strong anti-leukemic effect and it is a commonly used option for AML post-remission consolidation therapy.⁹ It especially improves the outcomes of patients with poor-risk or intermediate-risk disease.¹⁰ Although allo-HSCT can decrease the frequency of relapse and prolong survival, many patients still relapse post-transplant.¹ An increasing number of research focusing on post-transplant relapse have pointed out several factors including WBC count at diagnosis,¹¹ cytogenetic risk status,¹² initial induction response,¹³ and most importantly, minimal residual disease (MRD) status at the time of transplantation.¹⁴ In the MRD-negative patients who still experience relapse, genetic alterations might play an even bigger role in leukemogenesis and prognostication.¹⁵

The dedicator of cytokinesis 1 (DOCK) family is a class of the atypical Rho guanine nucleotide exchange factors (GEFs).¹⁶ As a major Rac GEF, DOCK proteins are involved in various cellular processes, such as cell adhesion, cell migration, actin cytoskeleton, and tumorigenesis.^17,18 Sze-Hwei Lee, et al, have discovered that high expression of DOCK1 implied poor prognosis in AML patients.¹⁹ However, it is yet to be determined whether the expression level of DOCK1 has prognostic value in AML patients undergoing allo-HSCT. In this study, we will focus on the prognostic significance of DOCK1 in a cohort of AML patients undergoing allo-HSCT and discuss its clinical implications.

Patients and methods

Patients

A group of 71 de novo AML patients (41 males, 30 females) from The Cancer Genome Atlas (TCGA) (https://cancergenome.nih.gov/) database, who had undergone allo-HSCT and had available DOCK1 expression data, were included in the study. TCGA database was originated from a single institution tissue banking protocol that was approved by the Washington University Human Studies Committee. All patients were between ages 18 and 88 and had previously untreated de novo AML. The samples were collected between November of 2001 and March of 2010. Clinical characteristics, as well as risk groups and the frequencies of known recurrently mutated genes, were highly representative of adult patients with de novo AML. The patients were treated in accordance with NCCN guidelines (www.nccn.org), with an emphasis on enrollment in therapeutic clinical trials whenever possible. Those patients with intermediate or unfavorable risk underwent allogeneic stem cell transplant if they were medically fit for transplantation, and if a suitably matched donor was available. Next-generation sequencing was utilized for detecting genetic mutations. Written informed consent was obtained from each patient approved by the Human Research Ethics Committee of Washington University. Clinical and molecular information at diagnosis were collected. All the data was publicly available on the website of the TCGA database.

Statistical analysis

Descriptive statistics were used to summarize demographics, clinical and molecular information. Continuous variables were described in the form of the median with range. The Mann–Whitney U test was used for comparing two groups of numeric variables, and chi-square analysis was applied to the comparison of categorical variables. The clinical endpoint of this study was overall survival (OS), which was defined as the time from diagnosis to death or was censored at the last follow-up. The OS rate was calculated using the Kaplan–Meier method, and the survival curves were plotted. The log-rank test was used to analyze significant differences between survival distributions. A cox proportional hazard model was applied to univariable and multivariable analyses to assess possible prognostic factors. The relapse-free survival (RFS) rate, defined as the time from the date of diagnosis to relapse, was also analyzed. The level of statistical significance was set at P<0.05 for all analyses. SPSS software version 20.0 and GraphPad Prism software version 6.0 were used for all the statistical analyses.

Results

Association of DOCK1 expression level with demographic characteristics and prognostic factors

The total of 71 patients was divided into two groups (DOCK1^high group and DOCK1^low group) based on the median DOCK1 expression level. High DOCK1 expressers were associated with older age (≥60, P=0.019), with an average age of 55 compared with 51 in DOCK1^low group (P=0.010). Patients with higher DOCK1 expression had a lower percentage of peripheral blood (PB) blasts (P=0.010) and fewer IDH1/2 mutations (P=0.010). FLT3-ITD appeared more frequently in DOCK1^high patients (P=0.051). All the patients with CEBPA double mutation were in DOCK1^low group while RUNX1 mutations were observed only in DOCK1^high group (P=0.005). There were no differences in other clinical factors including WBC count, FAB subtypes and karyotypes, as well as molecular prognostic features like TP53, MLL-PTD and NPM1 mutations. As for the ratios of relapse, no differences were found between the two groups. There was no difference between the two groups on the transplant status. The summary of the association between DOCK1 expression level and the demographic and prognostic factors is displayed in Table 1.

Table 1 Clinical and molecular characteristics of DOCK1high and DOCK1low patients

Higher expression level of DOCK1 indicated shorter OS in AML patients undergoing allo-HSCT

Univariate and multivariate analysis using cox proportional hazard models synthesized possible prognostic elements including DOCK1 expression level (high vs low), age (≥60 vs <60 years), WBC (≥30 vs <30×10⁹/L), risk stratification (poor vs non-poor) and genetic mutation like FLT3-ITD, NPM1, DNMT3A, CEBPA double mutation, IDH1 and RUNX1. The results are shown in Tables 2 and 3.

Table 2 Univariate analysis for OS

Table 3 Multivariate analysis for OS

Univariate analysis and Kaplan–Meier survival curve both revealed that patients undergoing allo-HSCT in DOCK1^high group had a shorter OS (Table 2, Figure 1, both P<0.001). Mutations in RUNX1 also had an adverse effect on OS (Table 2, P=0.024). The result of multivariate analysis further demonstrated the unfavorable effect of high DOCK1 expression (Table 3, P=0.010). Multivariate analysis also showed that patients who received a transplant in CR1/2 had longer OS (Table 3, P=0.043). RUNX1, FLT3-ITD and NPM1 were not correlated with the OS in AML patients undergoing allo-HSCT in our cohort. The results of univariate and multivariate analysis of RFS, as well as the Kaplan-Meier curve, were provided in Tables S1and S2 and Figure S1 in the Supplementary Appendix.

Figure 1 Kaplan–Meier curves of overall survival (OS). Patients in DOCK1high group have shorter OS than those in DOCK1low group.

Figure S1 Kaplan–Meier curves of relapse-free survival (RFS). There is no difference between DOCK1high group and DOCK1low group in the length of RFS.

Discussion

In our study, overexpression of DOCK1 can shorten OS in AML patients undergoing allo-HSCT, implying that allo-HSCT cannot eliminate the negative effect of DOCK1 overexpression. As an important prognostic marker, DOCK1 might be a potential therapeutic target for AML treatment.

Sze-Hwei Lee, et al¹⁹ showed that higher DOCK1 expression was associated with many other prognostic factors including intermediate-risk cytogenetics. As a curative-intent method, allo-HSCT has a well-documented anti-tumor effect especially on high-risk and intermediate-risk AML patients.²⁰ The prognostic influence of some prognostic mutations can be ameliorated by allo-HSCT.²¹ For example, Ma Y, et al concluded that allo-HSCT could efficiently reduce relapse and boost survival for patients with FLT3-ITD mutation.²² Yang Xu, et al have shown that allo-HSCT could prolong survival in cytogenetically normal AML.²³ Our study came to similar conclusions that allo-HSCT could overcome the adverse effect of some prognostic factors, such as FLT3-ITD, DNMT3A and RUNX1. However, the inferior outcome associated with DOCK1 overexpression could not be completely overcome by allo-HSCT. Though transplant status in CR1/2 prolongs survival to some extent, the statistical significance still brings the level of DOCK1 expression to our attention.

Consistent with the previous study,¹⁹ our study showed that patients with DOCK1 overexpression were older and had higher PB blast percentage at diagnosis. But there was no distinct relationship between DOCK1 expression and karyotype or cytogenetics, probably due to the implementation of allo-HSCT. We found a negative correlation between DOCK1 expression and frequencies of FLT3-ITD, RUNX1 mutations in our cohort. Since these were well-defined poor prognostic factors, the negative correlation could further imply that high DOCK1 expression could be an independent risk factor for AML outcome post-transplant.

As a GEF, DOCK1 plays multiple roles in physiological and pathological conditions. Researches have shown that DOCK1 regulates cell motility and tumor invasion in glioma cell by binding with engulfment and cell motility protein 1 (ELMO1).^24,25 Laurin, et al, reported that DOCK1 was a critical regulator in HER2-mediated breast cancer metastasis.¹⁸ In human colorectal cancer, DOCK1 was simulated by cortactin, which could promote cell migration and invasion.²⁶ Downregulation of DOCK1 may prevent epithelial–mesenchymal transition in bladder cancer.¹⁶ Furthermore, in vivo researches by Tajiri, et al. have demonstrated the effectiveness of targeting DOCK1 in the ras-driven cancer cell.²⁷ Hence, targeted therapy against DOCK1 deserves attention for future anti-cancer drug design. As demonstrated by the current study, the unfavorable effect of DOCK1 could not be eliminated by allo-HSCT, it would be worth investigating DOCK1-targeted therapy in AML patients for post-transplant maintenance.

Our study is limited by sample size and its retrospective nature. A larger cohort study is in need to confirm our findings. Further bench work exploring the role of DOCK1 in leukemogenesis are also imperative.

In conclusion, higher DOCK1 expression could predict a worse outcome in AML patients and its effect could not be overcome by allo-HSCT. DOCK1 is a potential anti-leukemia target for AML treatment.