CSRP2 transcript levels after consolidation therapy increase prognostic prediction ability in B-cell acute lymphoblastic leukemia

Quantification of measurable residual disease (MRD) correlates with the risk of leukemia recurrence in adults with B-cell acute lymphoblastic leukemia (ALL). However, it remains unknown whether collecting data on cysteine and glycine-rich protein 2 (CSRP2) transcript levels, after completing the second course of consolidation, improves prognosis prediction accuracy. A total of 204 subjects with B-cell ALL were tested for CSPR2 transcripts after completing the second course of consolidation using quantitative real-time polymerase chain reaction (qRT-PCR) and divided into high (N ═ 32) and low (N ═ 172) CSRP2 expression cohorts. In multivariable analyses, subjects with high expression of CSRP2 had a higher 5-year cumulative incidence of relapse (CIR) (hazard ratio [HR] ═ 2. 57, 95% confidence interval [CI] 1.38-4.76; P ═ 0.003), lower 5-year relapse-free survival (RFS) (HR ═ 3.22, 95% CI 1.75-5.93; P < 0.001), and overall survival (OS) (HR ═ 4.59, 95% CI 2.64-7.99; P < 0.001) in the whole cohort, as well as in the multi-parameter flow cytometry (MPFC) MRD-negative cohort (for CIR, HR ═ 2.70, 95% CI 1.19-6.12; for RFS, HR ═ 4.37, 95% CI 1.94-9.85; for OS, HR ═ 4.90, 95% CI 2.43-9.90; all P < 0.05). Prognostic analysis showed that allogeneic hematopoietic stem cell transplantation (allo-HSCT) could significantly improve the prognosis of patients with high CSRP2 expression (allo-HSCT vs chemotherapy: 5-year CIR, 52% vs 91%; RFS, 41% vs 9%; OS, 38% vs 20%; all P < 0.05). Our data indicate that incorporating data from CSPR2 transcript levels to the MRD-testing at the end of the second course of consolidation therapy enhances prognosis prediction accuracy in adults with B-cell ALL.


Introduction
In adults with B-cell acute lymphoblastic leukemia (ALL), which are completing initial therapy, there is a correlation between results of measurable residual disease (MRD) testing and subsequent risk of leukemia recurrence measured as cumulative incidence of relapse (CIR) [1][2][3][4].Most MRD-testing in adults with acute leukemia is based on multi-parameter flow cytometry (MPFC) detection of leukemia-associated immune phenotypes (LAIPs), quantitative polymerase chain reaction (qPCR) amplification-based methods detecting fusion genes, immunoglobulin or T-cell receptor (Ig/TCR) gene rearrangements, or next-generation sequencing (NGS) detecting leukemia-associated mutations [5][6][7][8].The patient scope and sensitivity of each method are different [9][10][11].However, some studies have shown that patients who are MRD positive by a PCR-based method but MRD negative by MPFC method are at increased risk for relapse compared with patients who are MRD negative with both methods [12][13][14].Consequently, additional methods with higher sensitivity for quantification of MRD and combined monitoring of multiple methods for MRD are needed to improve the ability of prognostic prediction.This would guide the refined risk stratification-based therapy, and ultimately, improve the long-term prognosis of adults with B-cell ALL.
The human cysteine and glycine-rich protein 2 gene (CSRP2) encodes the CSRP2 protein consisting of 193 amino acids with a molecular weight of about 21 KD [15].The CSRP2 protein contains two LIM domains with an inter-domain nuclear localization signal, which may function as a tool for the control of cell growth and differentiation [16,17].Hoffmann et al. [18,19] reported that CSRP2 expression was significantly upregulated in invasive breast cancer cells and its knockdown significantly reduced the invasive potential of human breast cancer cells in vitro.Tang et al. [20] reported that CSRP2 expression promoted pulmonary arteries smooth muscle cells (PASMCs) proliferation in vitro.We found that human CSRP2 transcript levels were upregulated in adults with B-cell ALL at the time of disease diagnosis, which correlated with a higher cumulative incidence relapse (CIR), especially in subjects with normal cytogenetics, and was associated with in vitro drug resistance [21].We investigated data obtained from 204 consecutive subjects with newly diagnosed B-cell ALL, after completing initial therapy, to determine if quantifying CSRP2 expression could be used to predict relapse.We found that MRD-testing at the end of the second course of consolidation therapy by CSRP2 transcript levels was an independent predictor of relapse and survival in multivariable analyses in subjects receiving subsequent maintenance chemotherapy or an allotransplant.

Subjects
A total of 1045 people with newly-diagnosed B-cell ALL were found at the Peking University Institute of Hematology from 2012 to 2019.Subjects younger than 15 years (N = 346) and/or subjects who received their initial therapy elsewhere (N = 373) were excluded from the study.In addition, 19 other subjects who did not achieve a complete hematological remission after 2 courses of induction chemotherapy were also excluded.Sixty-five of the remaining 307 subjects were excluded because of the relapse (N = 58) or death (N = 7) before completing the second consolidation therapy course.Another 9 subjects were excluded due to the discontinued follow-up, as well as 13 subjects with no available samples and 16 without complete MPFC data.The remaining 204 consecutive subjects, in the range of 15-69 years, were enrolled (Consolidated Standards of Reporting Trials diagram; Figure 1).
Diagnosis of B-cell ALL was based on World Health Organisation (WHO) 2016 criteria [22].Hematological complete remission was defined as bone marrow lymphoblasts < 5%, granulocyte concentration > 1.0 × 10E 9 /L, platelet concentration > 100 × 10E 9 /L, hemoglobin concentration > 100 g/L, no extra-medullary leukemia and no change in these criteria for > 1 month.Relapse was defined as the number of bone marrow lymphoblasts ≥ 5% at any site in subjects achieving hematological complete remission.
RNA extraction and synthesis of cellular DNA (cDNA) Mononuclear cells were isolated from bone marrow samples by Ficoll-Hypaque density gradient centrifugation (Solarbio Technology, Beijing, China) at diagnosis and after completing the second course of consolidation.Total cellular RNA was extracted using Trizol® kits (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions.Complementary DNA (cDNA) synthesis was done as described [24].

Measurement of relative CSRP2 transcript levels by qRT-PCR
Bone marrow samples at the end of the second course of consolidation therapy were analyzed by the relative transcript levels of CSRP2.TaqMan ® quantification was done using the ABI PRISM® 7500 FAST Sequence Detection System (Applied Biosystems, Foster City, CA, USA) with ABL1 as an internal control.The primer and probe sequence of CSRP2 and ABL1 were designed using Primer-Express software (Applied Biosystems) and displayed in Table S2.qRT-PCR was done as described [24].CSRP2 and ABL1 copy numbers were calculated as described in our previously published paper [21].

Ethical statement
The study was approved by the Ethics Committee of Peking University People's Hospital and all subjects have signed written informed consent consistent with the principles of the Helsinki Declaration.This trial has been registered in the Beijing Municipal Health Bureau Registration N 2007-1007 and in the Chinese Clinical Trial Registry [ChiCTR-OCH-10000940 and ChiCTR-OPC-14005546].

Statistical analysis
CIR was calculated as the interval from completing the second consolidation course to relapse, last follow-up, or withdrawal of consent.Cumulative incidences were estimated for relapse to accommodate competing risks.Relapse-free survival (RFS) was calculated from the completion of the second consolidation course to relapse, last follow-up, or withdrawal of consent.Overall survival (OS) was calculated as the interval from completing the second consolidation course to death, last follow-up, or withdrawal of consent.The threshold value to divide CSRP2 transcript levels into high and low cohorts was determined by the receiver operating characteristic (ROC) curve based on CIR data.Student's t-test and Mann-Whitney U tests were used to analyze normal continuous variables and non-normal continuous variables.Pearson chi-square or Fisher exact tests were used to evaluate categorical co-variates.The Bonferroni procedure was used to perform multiple comparisons.Survival functions were estimated by the Kaplan-Meier method and compared by the log-rank test.A Cox proportional hazard regression model was used to determine correlations among MRD defined by CSRP2 transcript level, RFS, and OS.A competing risk model was used to determine associations between CSRP2 transcript levels and CIR.Co-variates with P < 0.20 in univariable analyses were included in multivariable analysis.P < 0.05 in a 2-sided test was considered statistically significant.Analyses were performed by SAS version 9.4 (SAS Institute Inc., Cary, NC, USA), Graphpad Prism TM 9.0.0 (San Diego, California, USA), and R software package (version 4.0.3;http://www.r-project.org).A negative BCR::ABL1 at the end of the second course of consolidation therapy was defined as an individual ≥ 3 log reduction from the BCR::ABL1 transcript level at diagnosis [31].

Subject-and disease-related co-variates and outcomes
We studied 204 consecutive subjects who achieved a hematological complete remission after 1 (N = 192) or 2 (N = 12) courses of induction chemotherapy and remained in remission after 2 courses of consolidation therapy.In a preliminary analysis, there were no statistically significant differences in CIR, RFS, or OS between subjects requiring one or two induction chemotherapy courses to achieve a hematological complete remission and these cohorts were combined in subsequent analyses.The median follow-up of survivors was 31 months (interquartile range [IQR] 17-56 months).The median age of subjects was 34 years (IQR 24-46 years), and 111 were male.A total of 159 subjects (78%) received allotransplant, with a median of 3 months (IQR 2-4 months) after completing the second consolidation course.Forty-five others received maintenance chemotherapy only.Seventy subjects had a hematological relapse.The median interval from completing the second consolidation course to relapse was 13 months (IQR 5-25 months).Fifty-five subjects (27%) died of relapse (N = 39) or transplant-related mortality (N = 16).Details are displayed in Table 1.

Serial determinations of CSRP2 transcript levels
We studied serial determinations of CSRP2 transcript levels in bone marrow samples from eight subjects at diagnosis, in complete hematological remission, and at relapse.CSRP2 transcript levels in complete hematological remission were significantly lower compared with diagnosis or relapse samples (Figure 2A).In 3 subjects with long-term follow-up, we compared the results of CSRP2 testing with other MRD assays, including MPFC (N = 3), WT1 and BCR::ABL1 transcript levels (N = 1), and IKZF1 deletion (N = 1; Figures 2B-2D).CSRP2 transcript levels correlated well with clinical courses, as well as with other evaluated assays.
CSRP2 transcript levels after consolidation Subjects were divided into high (N = 32) and low (N = 172) cohorts based on a CSRP2 transcript at the end of the second course of consolidation therapy ≥ or < 0.93 percent of ABL1 transcript value determined by ROC curve based on CIR data.Clinical and laboratory co-variates were similar between cohorts except for white blood count (WBC) at diagnosis and MPFC-testing positivity or KMT2A rearrangement at the end of the second course of consolidation therapy (which was scored as MRD-positive; all P values < 0.05; Table 1).CSRP2 transcript levels were analyzed for correlations with results of MPFC-and BCR::ABL1-testing.The MPFC testing was positive in 17 subjects (53%) in the high CSRP2 transcript cohort vs 40 (23%; P = 0.001) in the low CSRP2 transcript cohort.BCR::ABL1 transcripts were detected in 7 subjects (22%) in the high CSRP2 transcript cohort vs 15 (15%; P = 0.43) in the low CSRP2 transcript cohort.One hundred and thirty-two (132/172, 77%) subjects were negative for MRD by MPFC-testing in the low CSRP2 transcript cohort and 17 (17/32, 53%) were positive for MRD by MPFC-testing in the high CSRP2 transcript cohort with a concordance of 73% (r = 0.82; P < 0.001; Figure 2E).In the high CSRP2 transcript cohort, 2 of 12 subjects who were BCR::ABL1positive at diagnosis became negative at the end of the second course of consolidation therapy compared with 26 of 74 in the low CSRP2 transcript cohort.Concordance for MRD-testing between BCR::ABL1 and CSRP2 transcripts was 65% (r = 0.98; P < 0.001; Figure 2F).

Discussion
Our data indicate that CSRP2 transcript levels, after the second course of consolidation therapy, are independently associated with 5-year CIR, RFS, and OS in adults with B-cell ALL receiving maintenance chemotherapy or an allotransplant.Providing data of CSRP2 transcript levels, in addition to results of MPFC-testing for MRD, improved relapse and survival prediction accuracy.
Included subjects with B-cell ALL were relatively young, with a median age of 34 years (IQR, 24-46 years), which probably reflected the transplant-related selection bias.The young age distribution may also explain the relatively few subjects with the BCR::ABL1 fusion gene.
There are several reasons why high CSRP2 level expression, after completing the second course of consolidation therapy, might correlate with an increased CIR.One is the indication of more residual leukemia cells compared with subjects with low CSRP2 expression.The second possibility relates to the mechanism of action of CSRP2 which favors cell proliferation, promotes cell-cycle progression, and inhibits apoptosis [18][19][20][21].These explanations are not mutually exclusive.
There are several important limitations to our study.First, it is a retrospective study.Second, sample size is relatively small and there was no external validation cohort to test the threshold value of CSRP2 transcript level.Third, our comparator MRD evaluation was MPFC rather than a molecular method, such as IGH rearrangement by NGS.Fourth, the relatively few subjects with BCR::ABL1 fusion genes limited the power of our analyses.Fifth, the assignment to post-consolidation therapy was not random.These limitations require further validation of our conclusions in larger prospective studies.

Conclusion
In summary, our data indicate that providing data about CSPR2 transcript level, in addition to the results of MPFC MRD-testing at the end of conventional therapy, improves relapse and survival prediction accuracy in adults with B-cell ALL regardless of subsequent therapy.

Figure 2 .
Figure 2. Correlation of CSRP2 transcript levels with clinical course and other MRD-tests.(A) Correlation between CSRP2 transcript level and clinical course in eight subjects at diagnosis, in complete hematological remission and at relapse; (B-D) Dynamic CSRP2 transcript levels in three subjects with long-term follow-up; (E and F) CSRP2 transcript levels with MPFC-MRD and BCR::ABL1 at the end of the second course of consolidation therapy.BM blasts: Bone marrow blasts; CSRP2: Cysteine and glycine-rich protein 2; MRD: Measurable residual disease; MPFC: Multi-parameter flow cytometry.

Figure 4 .Figure 5 .
Figure 4. Outcomes of CSRP2 transcript levels in negative MPFC-MRD subjects at the end of the second course of consolidation therapy.Cumulative incidence of relapse (A), relapse-free survival (B), and overall survival (C) were compared between subjects with high or low CSRP2 transcript levels.CON2: The second course of consolidation therapy; MPFC-MRD: Multi-parameter flow cytometry measurable residual disease; CSRP2: Cysteine and glycine-rich protein 2.

Table 3 .
Multivariable analyses of 5-year CIR, RFS, and OS in MPFC-negative subjects Detected at the end of the second course of consolidation therapy.CIR: Cumulative incidence of relapse; RFS: Relapse-free survival; OS: Overall survival; HR: Hazard ratio; CI: Confidence interval; MPFC: Multi-parameter flow cytometry; CSRP2: Cysteine and glycine-rich protein 2.