Circulating CXCL13 could be serving as a biomarker for chronic lymphocytic leukemia severity

Abstract

BACKGROUND:

Recent reports indicated the importance of chemotractant CXCL-13 in solid tumors and lymphoid malignancies. However, the prognostic value of the mentioned cytokines as biomarkers in chronic lymphocytic leukemia patient’s remains to be identified. Therefore; this study was designed in order to address the relation between CXCL-13 concentrations levels and markers of severity in CLL patients.

METHODS:

Our study included 150 CLL patients and 20 controls. Serum CXCL-13 was determined by ELISA for CLL patients at diagnosis as well as controls.

RESULTS:

The serum CXCL-13 levels were significantly higher in CLL patients as compared to controls. The high CXCL-13 concentration levels was significantly associated with high number of smudge cells; high LDH; high grade of Rai stage, short time to first treatment (TTT). Cox regression analysis was conducted for prediction of TTT, using age, gender, WBCs, smudge cells, CXCL-13, LDH, ZAP70, CD38, $\beta$ 2-microglobulin, Rai staging as covariates. High LDH, CXCL-13 and CD38% were significantly independent predictor for shorter TTT.

CONCLUSION:

High CXCL-13 serum levels at CLL diagnosis is correlated with other markers of disease activity; and could be served as biomarkers that predict CLL patient’s outcome.

Keywords

CXCL-13 CLL time to first time treatment

1. Introduction

Chemokines are a group of small molecular weight proteins has the power to act as chemoattractants, thereby functioning to induce the migration of nearby responding cells. These chemokines, together with extracellular mediators, including growth factors, are key modulators of inflammation by controlling complex interaction networks via an autocrine and a paracrine mechanisms. Dysregulation of chemokines has been reported in certain diseases including autoimmune disorders, CNS infections; and lymphoma. CXCL-13 serum levels was suggested as a biomarker for angio-immunoblastic lymphoma (AITL) [1, 2, 3, 4].

CXCL-13 is the ligand for CXCR-5, also known as the Burkitt lymphoma receptor-1 (BLR1), and the CXCL-13/CXCR-5 axis is necessary for B cell homing to lymph node follicles and for the production of immunoglobulin, which coordinates the humoral immunity of the body [5]. There is also growing evidence that CXCL-13 and CXCR-5 participate in the pathogenesis of multiple subtypes of lymphomas [6]. Malignant cells in follicular lymphoma were shown to express CXCR5, secrete CXCL13, and migrate in response to CXCL-13, suggesting that CXCL-13 recruits malignant B-cells to ectopic germinal centers and contributes to their development [7]. Elevated CXCL-13 levels have been reported in CNS B cell lymphomas [9, 10], patients with idiopathic multicentric Castleman Disease [10, 11], cutaneous B and T-cell lymphoproliferative disorders, intraocular lymphomas, and chronic lymphocytic leukemia [12].

CXCL-13 and its receptor CXCR-5 have emerged as key players of cancer initiation and progression. The identification of an autocrine and a paracrine interactions between the tumor microenvironment and cancer cells mediated by CXCL-13 highlights how autonomous and non-autonomous mechanisms contribute to the development of the cancer phenotype and the dissemination of cancer cells to metastatic sites [13].

The CXCL13 as chemoattractant organizes the cellular architecture of B-cell follicles and germinal centers. CXCL13 plasma concentrations function as a biomarker for normal germinal center activity and increases during adaptive immune response [14]. Previous studies reported that CXCL13 plasma levels was correlated with CLL B-cell expansion [12, 13, 14]. Therefore; we hypothesized that CXCL13 plasma concentrations could correlate with CLL disease activity and progression.

The aim of this study is to evaluate soluble CXCL-13 concentration as a biomarker for refining CLL severity.

2. Patients and methods

This study included 150 CLL patients and 20 controls. CXCL-13 was determined by ELISA at time of inclusion in parallel with well-established prognostic markers including time to first treatment (TTT). The diagnosis of CLL cases was according to WHO 2016. For each one of the patients the followings were carried out; blood counts; blood smear; immunophenotyping by flowcytometry using the following panel (CD5/CD19; CD23; FMC7, sIg, CD79b) as well as Cytogenetic findings (17pdel; 11qdel; 13pdel; trisomy 12) by FISH techniques. All included patients and controls had given informed consent and the study had been approved by local ethical committee at Mansoura University Oncology Centre. The patients were included in our study between January 2016 to January 2019.

Inclusion Criteria

CLL patients at diagnosis before start of therapy.

Exclusion Criteria

CLL patients Under therapy.

CLL patients combined with other malignancy.

2.1 Serum CXCL13 quantification by enzyme-linked immunosorbent assay (ELISA)

Blood samples were obtained at CLL diagnosis and from controls in blank tubes. Serum samples were separated and preserved in aliquot 1.5 ml in $-$ 80 ${{}^{\circ}}$ C until assay. The CXCL-13 ELISA procedure was done according to the manufacturer’s instructions (Quantikine ${}^{\text{\textregistered}}$ , R&D Systems, Minneapolis, MN). In brief, this sandwich enzyme immunoassay uses a mouse monoclonal antibody against human CXCL13 precoated onto a microplate as capture antibody. The standards, patients and control samples and biotin conjugated detection antibody were added to the wells subsequently, and washed with wash buffer. A mouse monoclonal antibody conjugated with horseradish peroxidase specifically directed against the human CXCL13 protein was added to the wells. Tetramethylbenzidine (TMB) was catalyzed by HRP to produce a blue color product that changed into yellow after adding acidic stop solution. The density of the developed yellow color is proportional to the CXCL13 amount of each sample captured in plate. Quantikine kit standards were used for construction of standard curves. We read the O.D. absorbance at 450 nm in a microplate reader, and then the concentrations of CXCL13 in patients samples and controls were calculated..

3. Statistical analysis

Categorical variables such as Rai stage risk, Cytogenetics abnormalities, CD38 and ZAP-70 expression were tabulated with frequency and percentage. While; continuous variables, such as white blood counts (WBC), LDH levels, $\beta$ 2-microglobulin levels, and the chemokine CXCL-13 were summarized with descriptive statistics, such as median and range. For analysis of associations, CXCL-13 median was used to dichotomize into groups of high ( $\geqslant$ median) and low ( $<$ median).

The Fisher’s exact test was used to evaluate the association between dichotomized CXCL-13 levels and categorical variables; the Mann Whitney U test or Kruskal-Wallis test was applied to compare between two groups or more than 2 groups regarding the continuous variables. Spearman rank correlations coefficients were used to evaluate the association between two continuous variables. The time to treatment (TTT) was defined as the time from sample collection to the date of the therapy; if patients have received any or censored on the last visit date. TTT was estimated by the method of Kaplan and Meier and assessed by the log-rank test to calculate the median TTT.

Cox models using univariate and multivariate were applied to evaluate the association of TTT with covariates. The clinical or laboratory variables which were significant in univariate analysis were retained in the multivariate analysis. The $P$ value $<$ 0.05 was considered significant.

4. Results

4.1 Clinical and laboratory characteristics in patient’s vs control

The present study was conducted on 150 CLL patients beside 20 healthy controls. Individuals in the CLL group were 105 males and 45 females; aged from (61.12 $\pm$ 10.1 years). Rai staging was categorized as low risk, intermediate risk and high risk (26.7%, 48.0% and 25.3% respectively). The most frequent cytogenetic finding was 13qdel which recorded in 53.3% of patients. The median WBCs, smudge cells, CXCL-13, and time to first treatment in CLL patients were presented in the Table 1. Positive ZAP-70, high CD38, and high $\beta$ 2-microglobulin were recorded in 44.7%, 34.0% and 35.3% respectively (Table 1).

Table 1
Patients characteristics

Parameter
Mean age (years)	61.12 $\pm$ 10.1
Gender
Male $n$ (%)	103 (68.6%)
Female $n$ (%)	47 (31.4%)
CLL patients $n$ (%)	150 (88.2%)
WBCS at diagnosis, median (range) ( $\times$ 10 ${}^{9}$ /L)	74.0 (32.0–320.0)
Smudge cell (%), median (range)	7.0 (1–27)
LDH at diagnosis, median (range)	337.5 (120–3200)
Rai staging
Low risk $n$ (%)	40 (26.7%)
Intermediate risk $n$ (%)	72 (48.0%)
High risk $n$ (%)	38 (25.3%)
ZAP 70%
Negative	83 (55.3%)
Positive	67 (44.7%)
CD38%
Negative	99 (66.0%)
Positive	51 (34.0%)
Beta2 microglobulin
Negative	97 (64.7%)
Positive	53 (35.3%)
Cytogenetic (FISH)
Normal	31 (20.7%)
17p del and 11q del	23 (15.3)
Trisomy 12	16 (10.7%)
13q del	80 (53.3%)
Median time to treatment (TTT)	16.0 (1–25)
Median CXCL13 at diagnosis	52.5 (0.1–1900.0)

Categorization of patient’s characteristics and laboratory parameters among CLL patients according to the patients age $<$ versus $\geqslant$ 60 years revealed that male patients was more frequent in CLL group with age $\geqslant$ 60 years. On the other hand; high $\beta$ 2-microglobulin were more frequently detected in age grou $p<$ 60 years. Otherwise; no significant differences regarding other data between the 2 age groups (Table 2).

Table 2

Comparison of patient’s characteristics and laboratory parameters among studied groups

Parameter		Age $\leqslant$ 60 ( $n=$ 62)		Age $>$ 60 ( $n=$ 88)		$P$
Gender	Male	33	53.2%	70	79.5%	0.001
Gender	Female	29	46.8%	18	20.5%	0.001
WBCS*	Median (Min-Max)	75.0 (32–320)		73.5 (35–310)		0.766
Smudge*	Median (Min-Max)	8.0 (3–27)		7.0 (1–24)		0.190
LDH*	Median (Min-Max)	390.0 (120–3200)		310.0 (128–2400)		0.644
CXCL13*	Median (Min-Max)	93.5 (0.2–1020)		48.50 (0.1–1900)		0.971
TTT*	Median (Min-Max)	14.5 (1–24)		16.0 (1–25)		0.169
Rai staging	Low risk	15	24.2%	25	28.4%	0.129
	Intermediate risk	26	41.9%	46	52.3%
	High risk	21	33.9%	17	19.3%
ZAP70	Negative	30	48.4%	53	60.2%	0.151
ZAP70	Positive	32	51.6%	35	39.8%	0.151
CD38	Negative	36	58.1%	63	71.6%	0.085
CD38	Positive	26	41.9%	25	28.4%	0.085
Beta2	Negative	32	51.6%	65	73.9%	0.005
Beta2	Positive	30	48.4%	23	26.1%	0.005
Fish	Normal	12	19.4%	19	21.6%	0.202
	17p del	4	6.5%	5	5.7%
	11q del	9	14.5%	5	5.7%
	Trisomy 12	9	14.5%	7	8.0%
	13q del	28	45.1%	52	59.1%

Mann-whitney tests*, Chi-square test, independent sample $T$ test** $P$ , between 2 groups. **Significant ( $P$ value $<$ 0.05).

Stratification of CLL laboratory findings according to Ria stages indicated that high WBCs, increased smudge cells, high LDH, and CXCL-13 ${}^{\text{high}}$ were frequently detected in high risk Rai staging as compared to other groups. On the other hand; TTT was significant shortened in high risk Rai stages as compared to other groups. Likewise; significant difference were detected ZAP-70, CD38, $\beta$ 2-microglobulin, and FISH cytogenetic findings among patients groups (Table 3).

Table 3

Comparison of patient’s characteristics and laboratory parameters among studied groups

Parameter		Low risk ( $n=$ 40)		Intermediate risk ( $n=$ 72)		High risk ( $n=$ 38)		$P$
Age**	Mean $\pm$ SD	62.2 $\pm$ 10.7		63.1 $\pm$ 9.07		56.4 $\pm$ 1.68		0.003
Gender	Male	28	70.0%	49	68.1%	26	68.4%	0.977
Gender	Female	12	30.0%	23	31.9%	12	31.6%	0.977
WBCS*	Median (Min-Max)	67.0 (43–115)		71.0 (32–122)		122.0 (62–320)		$<$ 0.001
Smudge*	Median (Min-Max)	7.0 (2–10)		6.0 (1–27)		14.5 (4–24)		$<$ 0.001
LDH*	Median (Min-Max)	221.0 (128–760)		293.5 (120–790)		1285.0 (340–3200)		$<$ 0.001
CXCL13*	Median (Min-Max)	88.5 (0.4–270)		13.5 (0.1–190)		655.0 (90.0–1900)		$<$ 0.001
TTT*	Median (Min-Max)	21.5 (7–24)		21.5 (3–25)		1.0 (1–3)		$<$ 0.001
ZAP70	Negative	37	92.5%	44	61.1%	2	5.3%	$<$ 0.001
ZAP70	Positive	3	7.5%	28	38.9%	36	94.7%	$<$ 0.001
CD38	Negative	38	95.0%	60	83.3%	1	2.6%	$<$ 0.001
CD38	Positive	2	5.0%	12	16.7%	37	97.4%	$<$ 0.001
Beta2	Negative	38	95.0%	57	79.2%	2	5.3%	$<$ 0.001
Beta2	Positive	2	5.0%	15	20.8%	36	94.7%	$<$ 0.001
Fish	Normal	31	77.5%	0	0.0%	0	0.0%	$<$ 0.001
	17p del and 11qdel	0	0.0%	0	0.0%	23	60.57%
	Trisomy 12	0	0.0%	1	1.4%	15	39.5%
	13q del	9	22.5%	71	98.6%	0	0.0%

Mann-whitney tests*, Chi-square test, ANOVA ** $P$ , between 3 groups. **Significant ( $P$ value $<$ 0.05).

4.2 Association between CXCL-13 levels and patients characteristics

Comparison of patient’s characteristics and laboratory parameters namely WBCs, smudge cells, and LDH were significantly elevated in CLL patients group with CXCL-13 ${}^{\text{high}}$ $\geqslant$ 52.5 as compared to those with CXCL-13 ${}^{\text{low}}$ $<$ 52.5. On the other hand TTT was significantly shorter in CLL group with CXCL-13 ${}^{\text{high}}$ $>$ 52.5 compared to other group. Moreover; there is significant difference as regard Rai staging, CD38, $\beta$ 2-microglobulin, and Cytogenetic finding among both groups (Table 4).

Table 4
comparison between Clincopathological paramaters in patients with high vs those with low CXCL13

Parameter		CXCL13 $<$ 52.5 ( $n=$ 75)		XCL13 $>$ 52.5 ( $n=$ 75)		$P$
Age**	Mean $\pm$ SD	62.5 $\pm$ 9.4		59.7 $\pm$ 10.7		0.091
Gender	Male	53	70.7%	50	66.7%	0.597
Gender	Female	22	29.3%	25	33.3%	0.597
WBCS*	Median (Min-Max)	70.0 (32–122)		77.0 (39–320)		$<$ 0.001
Smudge*	Median (Min-Max)	6.0 (1–27)		10.0 (3–26)		$<$ 0.001
LDH*	Median (Min-Max)	290.0 (120–790)		620.0 (128–3200)		0.001
TTT*	Median (Min-Max)	19.0 (4–25)		3.0 (1–24)		$<$ 0.001
Rai staging	Low risk	14	18.7%	26	34.7%	$<$ 0.001
	Intermediate risk	61	81.3%	11	14.7%
	High risk	0	0.0%	38	50.7%
ZAP70	Negative	47	62.7%	36	48.0%	0.071
ZAP70	Positive	28	37.3%	39	52.0%	0.071
CD38	Negative	63	84.0%	36	48.0%	$<$ 0.001
CD38	Positive	12	16.0%	39	52.0%	$<$ 0.001
Beta2	Negative	63	84.0%	34	45.3%	$<$ 0.001
Beta2	Positive	12	16.0%	41	54.7%	$<$ 0.001
Fish	Normal	13	17.3%	18	24.0%	$<$ 0.001
	17p del or 11qde	0	0.0%	23	30.0%
	Trisomy 12	0	0.0%	16	21.3%
	13q del	62	82.7%	18	24.0%

Mann-whitney tests*, Chi-square test, independent sample $T$ test** $P$ , between 2 groups. Significant ( $P$ value $<$ 0.05).

ROC analysis was conducted to identify the optimal Cut off levels for prediction of TTT. The analysis revealed that best cut-off value of CXCL-13 that could predict TTT was 65.5. The area under the curve (AUC) was 0.846 ( $p\leqslant$ 0.001) (Table 5; Fig. 1).

Table 5

Performance characteristics of CXCL13 for prediction of TTT

	AUC	SE	$p$	95% CI	Cut off	Sensitivity (%)	Specificity (%)
CXCL-13	0.846	0.033	$<$ 0.001	0.780–0.991	65.5	76.4%	72.6%

Figure 1.

ROC analysis was conducted to identify the optimal Cut off levels for prediction of shorter TTT. CXCL13 best cut-off value was 65.5. The area under the curve (AUC) was 0.846 ( $p\leqslant$ 0.001).

4.3 Predictive value of CXCL-13 serum concentrations to TTT

Of the 150 patients studied 110 progress and required treatment. Cox regression analysis was conducted to identify the best parameter that could predict TTT, using age, gender, WBCS, smudge cells count, CXCL-13, LDH, ZAP70, CD38, $\beta$ 2-microglobulin, Rai stages as covariates. High LDH HR 1.0 (CI 1.0-1.0; $p$ value 0.23), high CXCL13 serum levels HR 1.0 (CI 1.0-1.0; $P=$ 0.009) and CD38 positive HR 11.25 (CI 3,4 37,5; $P<$ 0.001) were significantly independent risk factor for shorter TTT.

4.4 Impact of CXCL-13 serum levels on CLL TTT was evaluated by Kaplan-Meier curve

To determine the differential impact of CXCL-13 serum levels on TTT, we categorized patients into 2 subgroups based on their CXCL-13 serum levels (CXCL-13 ${}^{\text{high}}$ / CXCL-13 ${}^{\text{low}}$ ) (Fig. 2). The cutoff level for this categorization was 52.5, CLL subgroup that was associated with higher CXCL-13 had a significantly shorter TTT as compared to those with low CXCL-13 serum levels. TTT estimates 82.7% at 12 months interval and 82.7% at 25 months interval in CXCL-13 $<$ 52.5, also TTT estimates 44.0% at 12 months and 44.0% at 25 months interval in CXCL-13 $>$ 52.5 with significant difference between 2 groups ( $P<$ 0.001) (Fig. 2).

Figure 2.

As regard CXCL13, TTT estimates 82.7% at 12 months interval and 82.7% at 25 months interval in CXCL13 $<$ 52.5, also TTT estimates 44.0% at 12 months and 44.0% at 25 months interval in CXCL13 $>$ 52.5 with significant difference between 2 groups ( $P<$ 0.001).

5. Discussion

CXCL-13/CXCR-5 axis plays central role in cancer development and progression. The interaction between cancer cells and its micro-environment through the an autocrine as well as a paracrine interactions which is mediated by CXCL-13 highlights different pathways that contribute to development and dissemination of cancer cells to the sites of metastasis [13]. CXCL-13 was identified as the most up regulated cytokine in plasma from patients with idiopathic multicentric Castleman disease [10].

The present study was conducted on 150 patients with CLL beside 20 healthy controls. Individuals in the study were 105 males and 45 females aged from (61.12 $\pm$ 10.1). Rai staging was diagnosed as low risk, intermediate risk and high risk (26.7%, 48.0% and 25.3% respectively). Most frequent cytogenetic was 13q del recorded in 53.3% of patients. Median WBCs, smudge cell, CXCL-13, and TTT in CLL patients presented in the table. Positive ZAP-70, CD38, $\beta$ 2-microglobulin were recorded in 44.7%, 34.0% and 35.3% respectively. These findings were similar to that reported by Sivina et al. [14].

Comparison of patient’s characteristics and laboratory parameters among studied subgroups, Male patients were frequently detected in group age $>$ 60 years. On the other hand; $\beta$ 2-microglobulin more significantly positive in group age $<$ 60 compared to group age $>$ 60 years. Otherwise no other significant differences could be detected. These findings were in parallel with that reported by Sivina et al. [14] as they compared these findings in CLL patients above 65 versus those less than 65 years

In the present study CXCL-13, WBCs and smudge cells were significantly higher in patients as compared to controls. Also, other laboratory parameters among studied groups namely, WBCs, smudge, LDH were significantly elevated in high risk Rai staging compared to other groups. On the other hand TTT was significant reduced in high risk Rai stage when compared to other groups. Also; there were significant difference as regard ZAP-70, CD38, $\beta$ 2-microglobulin, and FISH cytogenetic findings among groups. These findings were in agreement with that reported by Sivina et al. [14]. Our findings indicate that high CXCL-13 serum concentrations in CLL patients was detected in a subgroup of CLL patients that had active advanced stage disease, and shorter TTT. Our results suggested that CXCL-13 serum concentrations in CLL are mirror to the extent of disease activity as well as CLL proliferation [15].

Cox regression analysis was conducted for prediction of TTT, using age, gender, WBCs, smudge cells, CXCL-13, LDH, ZAP-70, CD38, $\beta$ 2-microglobulin, Rai stages as covariates. The multivariate analysis revealed that LDH, CXCL-13 and CD38 were significantly independent risk factor for shorter TTT. These findings were in parallel with the findings reported by Sivina et al. [14].

Survival analysis studies was evaluated by Kaplan Miere curve in order to address the impact of CXCL-13 $\geqslant$ 52.5vs $<$ 52.5) on TTT. TTT estimates 82.7% at 12 months interval and 82.7% at 25 months interval in CXCL-13 $<$ 52.5, also TTT estimates 44.0% at 12 months and 44.0% at 25 months interval in CXCL-13 $\geqslant$ 52.5 with significant difference between the 2 groups ( $P<$ 0.001). This finding indicates that high CXCL-13 is associated with shorter TTT in CLL patients. This is in agreement with that reported by single recent study [14].

6. Conclusion

Our findings indicated that CXCL-13 serum concentrations levels at CLL diagnosis is a valuable biomarkers for CLL disease staging as well as a marker for disease activity.

Funding statement

This work was not supported by a grant.

Author contributions

Salah Aref: Conception and supervision.

Ahmed Ramez: Interpretation and analysis of data.

Doaa Atia: Laboratory work; Preparation of Manuscript.

Tarek Abou Zeid: Interpretation and analysis of data.

Enas Gouda: Lab Work; Preparation of the manuscript; Revision for important intellectual.

Footnotes

Conflict of interest

The authors declare that there is no conflict of interest regarding the publication of this paper.

References

Dupuis

Boye

Martin

Copie-Bergman

Plonquet

Fabiani

et al., Expression of CXCL13 by neoplastic cells in angioimmunoblastic T-cell lymphoma (AITL): A new diagnostic marker providing evidence that AITL derives from follicular helper T cells, Am J Surg Pathol 30 (2006), 490–494.

Charo

I.F.

and Ransohoff

R.M.

, The many roles of chemokines and chemokine receptors in inflammation, N Engl J Med 354 (2006), 610–621.

von Hundelshausen

Agten

S.M.

Eckardt

Blanchet

Schmitt

M.M.

Ippel

et al., Chemokine interactome mapping enables tailored intervention in acute and chronic inflammation, Sci Transl Med 9 (2017).

Proost

Struyf

Van Damme

Fiten

Ugarte-Berzal

and Opdenakker

, Chemokine isoforms and processing in inflammation and immunity, J Autoimmun 85 (2017), 45–57.

Hussain

Adah

Tariq

Zhang

and Liu

, CXCL13/CXCR5 signaling axis in cancer, Life Sci 227 (2019), 175–186.

Widney

D.P.

Breen

E.C.

Boscardin

W.J.

et al., Serum levels of the homeostatic B cell chemokine, CXCL13, are elevated during HIV infection, J Interferon Cytokine Res 25 (2005), 702–706.

Husson

Freedman

A.S.

Cardoso

A.A.

et al., CXCL13 (BCA-1) is produced by follicular lymphoma cells: Role in the accumulation of malignant B cells, Br J Haematol 119 (2002), 492–495.

Fischer

Korfel

Pfeiffer

et al., CXCL13 and CXCL12 in central nervous system lymphoma patients, Clin Cancer Res 15 (2009), 5968–5973.

Rubenstein

J.L.

Wong

V.S.

Kadoch

et al., CXCL13 plus interleukin 10 is highly specific for the diagnosis of CNS lymphoma, Blood 121 (2013), 4740–4748.

10.

Pierson

S.K.

Stonestrom

A.J.

Shilling

et al., Plasma proteomics identifies a ‘chemokine storm’ in idiopathic multicentric Castleman disease, Am J Hematol 93 (2018), 902–912.

11.

Smith

J.R.

Braziel

R.M.

Paoletti

et al., Expression of B-cell-attracting chemokine 1 (CXCL13) by malignant lymphocytes and vascular endothelium in primary central nervous system lymphoma, Blood 101 (2003), 815–821.

12.

Bürkle

Niedermeier

Schmitt-Gräff

et al., Overexpression of the CXCR5 chemokine receptor, and its ligand, CXCL13 in B-cell chronic lymphocytic leukemia, Blood 110 (2007), 3316–3325.

13.

Kazanietz

Durando

and Cooke

, CXCL13 and its receptor CXCR5 in cancer: Inflammation, immune response, and beyond, Frontire in Endocrinology (2019), REVIEW published: doi: 10.3389/fendo.2019.00471.

14.

Sivina

Xiao

Kim

Vaca

Chen

Keating

et al., CXCL13 plasma levels function as a biomarker for disease activity in patients with chronic lymphocytic leukemia, Leukemia (2020). doi: 10.1038/s41375-020-01063-7.

15.

Herndon

T.M.