CD9 expression indicates a poor outcome in acute lymphoblastic leukemia

Abstract

OBJECTIVE:

We undertook a single-center retrospective study to determine the relationship between CD9 and acute lymphoblastic leukemia (ALL).

MATERIALS AND METHODS:

In total, 112 newly diagnosed patients in our center were enrolled in the study. Their clinical information was collected and the patients werefollowed over the course of the study. Flow cytometry was used to detect the expression of CD9.

RESULTS:

CD9 expression was more common in B cell acute lymphoblastic leukemia (B-ALL) and patients $>$ 40 years old. CD9-positive patients exhibited a higher BCR-ABL fusion gene positive rate and higher neutrophil counts than CD9 negative patients ( $P=$ 0.004 and $P=$ 0.004, respectively). Response to induction chemotherapy was not dependent on CD9 expression. CD9-positive patients had a lower 2-year overall survival rate than CD9-negative patients.

CONCLUSION:

CD9 expression predicts some clinical characteristics and indicates an unfavorable prognosis in ALL patients.

1. Introduction

Acute lymphoblastic leukemia (ALL) is a common hematological malignancy in adolescents and adults. Although new laboratory examinations and drugs have been beneficial to patients, the prognosis of ALL remains dissatisfactory. Immunophenotyping analysis with flow cytometry is a vital tool for ALL diagnosis and minimal residual disease monitoring. CD9 is usually detected as an immunophenotyping marker in ALL. CD9 is a cellular surface glycoprotein and belongs to the tetraspanin superfamily, which consists of at least 33 conserved proteins including CD9, CD63, CD81, and CD37. These proteins contain four transmembrane domains and play crucial roles in cellular growth, motility, and signal transduction [1]. As a cell membrane glycoprotein, CD9 interacts with other proteins to transduce signals, promoting apoptosis and inhibiting metastasis [2]. The expression of CD9 has been shown to be linked with the prognosis of a number of solid tumors, such as colorectal and gastric cancer, and hematological malignancies including follicular lymphoma and multiple myeloma [3]. Precursor B cells in bone marrow have high CD9 expression while mature B cells show downregulation of CD9. However, CD9 is re-expressed in plasma cells [8]. CD9 can also moderate the properties of cancer stem cells and the migration and engraftment of B-ALL cells in the bone marrow [9]. These reports suggest that CD9 may participate in the development and differentiation of B cells and the development of the lymphoblastic malignancies. Up until now, little has been reported on the relationship between CD9 and ALL with respect to clinical characteristics and overall outcome. In this study, we collected the clinical data of newly diagnosed ALL patients and analyzed the associations between CD9 expression and prognosis as well as other clinical characteristics.

Table 1

Basic characteristics stratified by CD9

	CD9 ${}^{+}$ ( $n=$ 68)	CD9 ${}^{-}$ ( $n=$ 44)	$P$
Gender (male:female)	27:41	21:23	0.439
Age (%)
Older ( $>$ 40 years)	60.29	34.09	0.011#
Young ( $⩽$ 40 years)	39.71	65.91
Immunophenotyping (%)
B-ALL	97.06	59.09	0.000#
T-ALL	2.94	40.91
BCR-ABL ${}^{+}$ (%)	30.88	6.38	0.004#
Mutant genes (%)	5.88	13.63	0.187
Abnormal karyotypes (%)	29.41	15.91	0.103
Fever	51.47	45.45	0.556
Enlarged lymph nodes (%)	20.59	36.36	0.066
Sternal tenderness (%)	20.59	15.91	0.625
Bleeding (%)	41.18	38.63	0.845
CD34 ${}^{+}$ (%)	72.06	61.36	0.301
BM blasts (Mean $\pm$ SD)	82.16 $\pm$ 12.65	78.35 $\pm$ 13.05	0.20
WBC counts (Mean $\pm$ SD)	61.72 $\pm$ 107.06	34.73 $\pm$ 123.25	0.301
Neutrophil (Mean $\pm$ SD)	5.42 $\pm$ 6.95	2.44 $\pm$ 2.57	0.004*
Lymphocyte (Mean $\pm$ SD)	11.07 $\pm$ 45.77	4.45 $\pm$ 11.36	0.469
Hb (Mean $\pm$ SD)	85.42 $\pm$ 34.75	75.35 $\pm$ 28.91	0.195
PLT (Mean $\pm$ SD)	62.19 $\pm$ 164.59	74.15 $\pm$ 74.50	0.723
LDH (Mean $\pm$ SD)	679.36 $\pm$ 835.71	547.04 $\pm$ 902.74	0.507

#: ${\chi }^2$ test; *: T test.

2. Materials and methods

We investigated newly diagnosed ALL patients in our center from January 2013 to October 2016. Patients without complete clinical information or laboratory test results were excluded. In total, 112 patients were enrolled. All patients were diagnosed according to World Health Organization 2008 criteria for ALL with morphology, immunology, cytogenetics, and molecular biology [10]. Our study was approved by the Institutional Review Board of Shengjing Hospital of China Medical University. CD9 expression was detected withflow cytometry using fluorochrome-conjugated antibodies. CD9 was considered to be strongly positive when it was found on more than 80% of all the blasts, 20% to 80% as weakly positive, and $<$ 20% as negative. Fusion genes were detected by polymerase chain reaction and karyotypes were examined by fluorescent in situ hybridization.

Table 2

Factors affecting ORR stratified by CD9

	ORR rate of	ORR rate of	$P$
	CD9 ${}^{+}$ group	CD9 ${}^{-}$ group
Overall	64.15	52.94	0.466
Gender
Male	55.00	71.43	0.477
Female	81.82	65.00	0.200
Age
$>$ 40 years	72.00	84.62	0.456
$⩽40$ years	83.33	64.00	0.196
Immunophenotype
B-ALL	70.59	70.00	1.000
T-ALL	100.00	64.29	1.000
WBC counts
High	62.50	71.43	0.671
Low	76.32	66.67	0.414
BCR-ABL
$+$	82.35	100.00	0.558
$-$	66.67	62.07	0.796
Karyotypes
Low-risk	78.38	64.29	0.267
High-risk	56.25	83.33	0.351

Of the 112 patients enrolled in the study, 25 did not receive regular therapy because of economic problems or physical state. The remaining 87 patients received inductionchemotherapy according to Chinese guidelines for the diagnosis and treatment of ALL [11]. After induction treatment, we evaluated the curative effect with complete response (CR), CR with incomplete blood count recovery (CRi), refractory disease, progressive disease, and relapsed disease. Overall response rate (ORR) is equal to CR and CRi. BCR-ABL-positive patients who received chemotherapy were also treated with tyrosine kinase inhibitors (TKIs). All of the evaluation criteria were based on the Chinese guidelines for ALL. Progression free survival (PFS) was calculated from the date of CR achievement to any kind of disease relapse. Overall survival (OS) was defined as the period from diagnosis to death. Early death meant patients died within 30 days of diagnosis.

2.1 Statistical analysis

All data were performed with SPSS version 19.0 (IBM, Armonk, NY, USA). Count data are presented as percentages (%) whereas measurement data are presented as mean values $\pm$ standard deviation (SD). Student’s $t$ -test was used to detect the differences in measurement data and ${\chi }^2$ test or Fisher’s exact test was used for count data. The OS rates were calculated by Kaplan–Meier curves. Cox regression was used to deal with univariate and multivariate analysis of OS. $P$ values less than 0.05 were considered to indicate statistical significance.

3. Results

3.1 Baseline characteristics of ALL patients

Clinical characteristics of the 112 patients (48 men and 64 women) are shown in Table 1. Patient age ranged from 14 to 75 years. The median age was 40.00 years old and the average age was 39.46 years old. Patients $⩽$ 40 years were categorized asthe younger group and patients $>$ 40 years were categorized as the older group. In total, 63 of the 112 patients were strongly positive for CD9 and 5 of the 112 patients were weakly positive. Thus, the overall expression rate of CD9 was 60.71%. CD9 expression was more often seen in B-ALL patients ( $P<$ 0.001) and older patients ( $P=$ 0.011). Gender did not influence CD9 expression. The CD9-positive (CD9 ${}^{+}$ ) group exhibited higher BCR-ABL fusion gene expression and higher neutrophil counts than the CD9 negative (CD9 ${}^{-}$ ) group ( $P=$ 0.004 and $P=$ 0.004, respectively).

3.2 CD9 expression does not affect the response to induction chemotherapy

The response to induction chemotherapy is an independent risk factor ofALL prognoses. For various reasons, only 87 patients underwent induction chemotherapy. In total, 59.77% of patients achieved CR and 10.34% achieved Cri, and thus, the ORR was 70.11%. In addition, 29.89% of patients failed to reach ORR after the induction period. We found no significant difference in ORR between the CD9 ${}^{+}$ and CD9 ${}^{-}$ groups. We next analyzed the relationship between CD9 and the induction response in gender, age, immunophenotype, white blood cell (WBC) counts, fusion genes, and karyotypes but found CD9 did not affect induction chemotherapy response. Details concerning the CR rates of relevant groups are displayed in Table 2.

Table 3

Univariate and multivariate analysis for OS in patients undergoing induction chemotherapy

	$P$ value	HR	95% CI
Univariate analysis
CD9	0.014*	2.165	1.171–4.002
Gender	0.796	0.930	0.583–1.609
Age	0.042*	1.758	1.020–3.028
Immunophenotypes	0.523	0.7971	0.397–1.599
Induction response	0.000*	6.440	3.463–11.979
Karyotypes	0.317	1.366	0.741–2.517
BCR-ABL	0.748	0.899	0.471–1.716
BM blasts	0.977	1.000	0.978–1.022
WBC counts	0.244	1.432	0.783–2.620
Multivariate analysis
CD9	0.018*	2.206	1.143–4.256
Age	0.678	0.158	0.085–0.292
Induction response	0.000*	6.429	3.439–12.019

*: Cox regression.

Figure 1.

(A) Overall survival (OS) curve of 112 patients stratified by CD9. (B) OS curve of 112 patients stratified by CD9 and age. (C) OS curve of 87 patients receiving induction chemotherapy stratified by CD9. (D) OS curve of 87 patients receiving induction chemotherapy stratified by CD9 and chemotherapy response. (E) OS curve of 71 B-ALL patients receiving induction chemotherapy stratified by CD9 and BCR-ABL. (F) OS curve of 24 BCR-ABL ${}^{+}$ B-ALL patients receiving induction chemotherapy stratified by CD9.

3.3 CD9 expression indicates an unfavorable outcome in ALL patients

In our study, the median OS of 112 ALL patients was 6 months (95% confidence interval [CI]: 7.3–11.2 months). The 2-year OS rate of the CD9 ${}^{+}$ group was lower than the CD9 ${}^{-}$ group (7.35% vs. 15.91%, $P=$ 0.041; Fig. 1A). The early death rate of the 112 patients was 16.07% and expression of CD9 had no impact on early death rate (19.12% vs. 11.36%, $P=$ 0.307). CD9 expression had no effect on the 2-year PFS rate in which the PFS rate of CD9 ${}^{+}$ was 4.41% and CD9 ${}^{-}$ was 13.64% ( $P=$ 0.416). Given the significance of the induction chemotherapy response in prognosis, we then analyzed the survival of the 87 patients who received induction therapy. Among those patients, the median OS was 7 months (95% CI: 8.9–13.6 months). The 2-year OS rate of the CD9 ${}^{+}$ group was 9.43% compared with 20.59% in the CD9 ${}^{-}$ group ( $P=$ 0.036; Fig. 1B). Expression of CD9 had no impact on PFS (5.66% vs 17.65%, $P=$ 0.074). In univariate analysis, CD9 expression, age, and induction response reached statistical significance ( $P=$ 0.014, $P=$ 0.042, $P<$ 0.001; Table 3). Nevertheless, in multivariate analysis, only CD9 and the induction response had a significant effect on OS (Table 3). From the survival curve we could see that the CD9 ${}^{-}$ and CR groups had the best OS (Fig. 1C). Next, we analyzed the effect of CD9 on OS in the CR and NR groups. We found that CD9 expressionresulted in a poor OS in the CR group ( $P=$ 0.001) but did not influence OS in the NR group ( $P=$ 0.342). To better understandwhy age failed to reach statistical significance, we analyzed the effect of CD9 expression and the response to induction chemotherapy on OS in young and older patients. CD9 expression did not influence OS in young patients ( $P=$ 0.874; Fig. 1D) but resulted in a shorter OS time in older patients ( $P=$ 0.034; Fig. 1D). The BCR-ABL fusion gene is a crucial factor in ALL. We analyzed the effects of CD9 and BCR-ABL on OS in B-ALL patients undergoing induction chemotherapy ( $n=$ 71), but found it failed to reach statistical significance ( $P=$ 0.089; Fig. 1E). Among BCR-ABL ${}^{+}$ B-ALL patients ( $n=$ 24), CD9 ${}^{+}$ patients exhibited worse OS than CD9 ${}^{-}$ patients ( $P=$ 0.034; Fig. 1F). Overall, we found that CD9 predicts a poor prognosis for ALL patients.

4. Discussion

There have been a few reports regarding the relationshipof CD9 and ALL. During B cell development, the expression of CD9 fluctuates. CD9 shows upregulation in precursor B cells, downregulation in mature B cells, and re-expression in plasma cells [8]. Gandemer et al. [12] found that CD9 expression was lower in TEL/AML1 ${}^{+}$ ALL patients than in TEL/AML1 ${}^{-}$ patients. Aoki et al. [13] reported that leukemia-initiating cellsin which MLL-AF4, MLL-AF9, and MLL-ENL fusion genes were apparent exhibited higher expression of CD9 than normal hematopoietic stem cells. Nishida et al. [14] and Yamazaki et al. [9] considered CD9 as a marker of cancer stem cells and suggested it could regulate cancer stem cell function. Leung et al. [15]demonstrated that CD9 was important for the migration, adhesion, and homing of CD34 ${}^{+}$ HSCs. Arnaud et al. [16] showed that CD9 promoted RAC1 activation and enhanced the migration of ALL cells in the bone marrow. Our data showed that the CD9 positive rate was 60.71%. The CD9 ${}^{+}$ group had higher BCR-ABL fusion gene expression and higher neutrophil counts than the CD9 ${}^{-}$ group. The relationship between CD9 and the BCR-ABL fusion gene has not been well studied. In this study, in BCR-ABL ${}^{+}$ patients, the CD9 ${}^{+}$ rate was shown to be 87.5%. CD9 is a target of Bruton’s tyrosine kinase (Btk) and can be downregulated when Btk is missing [17]. Btk is downstream of BCR-ABL and participates in BCR-ABL-mediated B-cell transformation and resistance to imatinib [18]. Considering this, we suggest that CD9 expression may be linked to BCR-ABL signaling. The expression of CD9 was extremely different betweenB-ALL and T-ALL but CD9 expression had no difference among the different classifications of B-ALL. Therefore, we have discovered some influences of CD9 on ALL and provided possibilities for future exploration of the role of CD9 in ALL.

Previous studies have reported that CD9 plays different roles in different types of tumors. In some solid tumors, such as colorectal carcinoma and gastric cancer, expression of CD9 indicates a better prognosis. In breast cancer cells, however, expression of CD9 promotes invasiveness and metastasis [19]. In follicular lymphoma, expression of CD9 indicates a longer PFS and higher bone marrow infiltration. In multiple myeloma, CD9 expression is correlated with $\beta$ 2 microglobulin and hemoglobin levels and indicates a satisfying prognosis [20]. With respect to leukemia, it had been reported that patients that express CD9 had a lower CR rate than CD9-absent patients in AML [21]. In our study, CD9 expression did not influence the curative effect of induction chemotherapy, even when taking other factors into consideration. Response to induction chemotherapy is a pivotal factor for the prognosis of ALL patients. CD9 expression, age, and response to induction chemotherapy were risk factors for OS but when considering these three factors in multivariate analysis, age ultimately lost significance. CD9 expression showed poor outcome in CR patients but was not significantin NR patients. This might be caused by the short OS time of the NR group with the median OS time of 1.5 months, which is much shorter than the entire median OS time. We can see that the induction response is a significant factor for prognosis. Age is another master factor for prognosis. CD9 status could distinguish OS in the older group but failed in the young group. CD9 ${}^{+}$ and older group exhibited the most dissatisfactory outcome and the CD9 ${}^{+}$ and young group was slightly better. In the CD9 ${}^{-}$ groups, the outcome was more satisfactory. All of these results suggest that CD9 is a risk factor for ALL patients.

Our study was a single-center analysis and the total number of patients, as well as the number of patients who underwent regular treatment, was not sufficient. Moreover, the BCR-ABL fusion gene and mutant genes, which are considered risk factors in the Chinese guidelines, did not affect OS. We believe that this might be the result of the low positive rates of corresponding elements. We also did not distinguish the immunophenotypes not only because they failed to reach statistical significance but we also wished to identify a general factor to evaluate ALL prognoses. The CD9 ${}^{-}$ and older group appeared to exhibitthe highest OS, however there were only nine patients in this group and only one died before follow-up was halted. In the future, we will optimize our study and further investigate the mechanisms of CD9 function in ALL cells.

5. Conclusions

Our study demonstrates that CD9 ${}^{+}$ ALL patients have higher BCR-ABL fusion gene expression rate and higher neutrophil counts. CD9 has no effect on the patients’ response to induction chemotherapy. Furthermore, CD9 ${}^{+}$ patients possess a lower 2-year OS rate and CD9 expression is a risk factor for OS.