Early Recovery of Circulating Immature B Cells in B-Lymphoblastic Leukemia Patients after CD19 Targeted CAR T Cell Therapy: A Pitfall for Minimal Residual Disease Detection

Abstract

Background:

CD19-targeted chimeric-antigen receptor-modified T-cells (CAR-T) are promising in the treatment of refractory B-lymphoblastic leukemia (B-ALL). Minimal residual disease (MRD) detection by multicolor flow cytometry (FCM) is critical to distinguish B-ALL MRD from regenerating, non-neoplastic B-cell populations.

Methods:

FCM was performed on samples from 9 patients with B-ALL treated with CAR-T.

Results:

All 9 patients showed response to CAR-T. Additionally, FCM revealed circulating CD10 + B cells, potentially mimicking MRD. Circulating CD10+ B-cells were detected in blood from 3 days to 3 months after CAR-T, comprising 73% (median) of B-cells (52–83%, 95%CI). They expressed CD19, CD10, CD20, bright CD9, CD22, CD24, moderate CD38 and dim CD58, but were CD34 (−), with bright CD45 and polyclonal surface light chain immunoglobulin (sIg) expression. A similar CD10 + B-cell subpopulation was detected by marrow FCM, amidst abundant B-cell precursors.

Conclusions:

These circulating CD10 + B-cells are compatible with immature B-cells, and are a reflection of B-cell recovery within the marrow. They are immunophenotypically distinguishable from residual B-ALL. Expression of light chain sIg and key surface antigens characterizing regenerating B-cell precursors can distinguish immature B-cells from B-ALL MRD and prevent misdiagnosis. © 2017 International Clinical Cytometry Society

Remarkable progress has been made in the treatment of B-lymphoblastic leukemia (B-ALL), the most common childhood malignancy, with 5-year survival approaching and even exceeding 90%. However, relapsed disease remains a leading cause of B-ALL related death in children and outcomes for adults with B-ALL remain poor (1,2). Recently, adoptive immunotherapy using T-cells genetically engineered to express a chimeric antigen receptor (CAR-T) targeting CD19, a cell surface molecule present in virtually all B-ALL, have been used to treat patients with refractory/relapsed B-ALL, with promising results (3–7). Monitoring of minimal residual disease (MRD) by multicolor flow cytometric (FCM) is routine in the management of childhood B-ALL and is also informative in adult B-ALL (8–10). B-ALL MRD at the end of induction therapy, consolidation therapy, and prior to allogeneic hematopoietic stem-cell transplantation (HSCT) is associated with significantly worse outcome (11–13). MRD assessments are prognostically valuable for risk stratification, ranging from significant treatment reduction to mild or strong intensification (10,14). We recently encountered a recovering population of circulating CD10+ B cells in patients with B-ALL after CAR-T that may confound MRD detection by FCM; therefore, we examined this population in the context of regenerating marrow precursors as well as the patients’ B-ALL blast phenotype at diagnosis and relapse.

MATERIALS AND METHODS

Patients

The 9 B-ALL patients described were enrolled in clinical trial (Phase I Study of anti-CD19 Chimeric Antigen Receptor therapy for children and young adults with relapsed/refractory ALL; NCT01593696, registered with ClinicalTrials.gov) from 2013–2015 with oversight by NIH Institutional Review Board, and received CD19 CAR-T cells as reported (7).

Flow Cytometry

PB and BM were collected pre- and post- CAR-T, and processed for 8-color FCM within 24 h of collection with cocktailed antibodies (see Supporting information), based on Children’s Oncology Group guidelines (14), and modification of the Cherian et al. approach (15). FCM procedure (16) included NH₄Cl whole blood lysis, phosphate-buffered saline wash and antibody incubation. Cells were then washed, pelleted, fixed in 1% formalin and stored at 4°C for <12 h prior to acquisition. A target of 1,000,000 cells per tube/cocktail was acquired on a FACSCanto II flow cytometer (BD Bioscience), and analyzed with FCSExpress 4 (DeNovo Software, Glensdale CA). Mononuclear cells were gated by light scatter with antigen back-gating to verify relevant populations. Normal hematopoietic cells within the specimens served as internal positive/negative controls.

Statistics

Student t-tests were performed using Prism 6 (Graph-Pad Software Inc., La Jolla, CA). Two-sided P values are reported with a type I error rate of 5% and a P < 0.05 set for significance.

RESULTS

Patients

Of the 8 males and 1 female, the median age was 15 years (range: 5–25) and included both primary refractory (without achieving disease-negative status) and relapsed B-ALL. Prior to enrollment, all had received multiple cycles of chemotherapy; a subset had received prior allogeneic HSCT. Immediately prior to CAR-T, 7/9 had FCM-detectable disease only (no morphologically-detectable marrow disease), while 2/9 had >5% B-ALL blasts in the BM. At Day 28 post CAR-T, 9/9 patients were FCM MRD negative in BM.

FCM Prior to CAR-T

The immunophenotype (IP) of each patient’s B-ALL in bone marrow prior to CAR-T is described in Table 1. For 8/9 patients, the IP data was generated in-house; due to low level disease, full IP characterization of Patient 7 could not be performed, and was supplemented by flow cytometry data reported from an outside institution. B-ALL blasts exhibited low SSC and all expressed CD19 and CD22. CD24 was positive on 6/6 patients, where testing was informative. CD34 expression was detected in all patients, with partial expression noted in 1 patient, and dim to moderate expression noted in another. CD38 expression was low (either dim or negative) for all patients. Aberrant CD33 expression was observed on 3/8 patients. No aberrant CD13 expression was detected in 8/8 patients where testing was informative. Interestingly, 6 patients exhibited some degree of CD20 expression on blasts. CD45 was dim to negative in 8 cases and moderate in 1 case. CD10 was bright in 6 of 8 CD10 positive cases and CD58 was bright in 5 of 8 CD58 positive cases. CD9 was positive in 7 cases and dim to negative in 2 cases. All cases lacked light chain sIg expression.

Table 1.

Immunophenotype of B-ALL Prior to CD19CAR-T Therapy

Patient	CD9	CD10	CD19	CD20	CD22	CD24	CD34	CD38	CD45	CD58	CD13	CD33	slg
1	+	br	+	(−)	+	+	br	dim	dim to (−)	br	(−)	(−)	(−)
2	+	br	+	(−)	+	+	br	dim to (−)	dim to (−)	+	(−)	(−)	(−)
3	Br	dim	br	dim	+	NR	dim to moderate (+)	dim	dim to (−)	br	(−)	dim	(−)
4	dim to (−)	br	+	+	+	+	br	dim to (−)	dim	+	(−)	+	(−)
5	(−)	br	+	br	+	NR	Partially (+)	dim	dim	(−)	(−)	(−)	(−)
6	+	Partial (+)	+	Partial (+)	+	NR	br	dim	dim	br	(−)	+	(−)
7a	Br	(−)	+	NR	+	+	br	(−)	Moderate (+)	br	NR	NR	(−)
8	+	br	+	dim to moderate (+)	+	+	+	(−)	dim	br	(−)	(−)	(−)
9	+	br	+	+	+	+	+	dim	dim to (−)	+	(−)	+	(−)

Abbreviations: CAR-T=chimeric antigen receptor T-cell; + = positive; (−) = negative; br = bright positive; NR = not reported.

^aReported from outside institution.

Circulating Immature B-Cells

After CAR-T, we observed circulating CD19 + CD20+ B-cells expressing CD10 in these 9 patients, all of whom demonstrated response to CAR-T, consistent with immature B-cells, according to the 8-stage B-cell differentiation nomenclature ultilized by Noordzu et al. (17). The immature B-cells were detected as early as 3 days, and up to 3 months after CAR-T, comprising 73% (median) of B-cells (52%−83%, 95% CI, Table 2). Figure 1C–E illustrates the significant population of immature B-cells in a patient post CAR-T. In contrast, a representative PB and BM sample (Fig. 1A and B) from a disease-free Patient (3 years post treatment, unrelated clinical protocol) are shown for comparison. The circulating B cells were pre-dominantly mature (i.e., CD10 negative) with only rare (< 1%) immature B cells (i.e., CD10 positive) (Fig. 1B); the concurrent aspirate FCM (Fig. 1A) demonstrated progressive maturation of regenerating CD19+ B-cell subpopulations. The majority of the B-cells in BM exhibited CD10 and variable CD20 expression, typical for regenerating B-cell precursors of various stages (e.g., Pre-B-I, Pre B-II, Immature B, Mature B). The circulating immature B-cells from the patient’s PB post CAR-T resembled a normal population of regenerating B-cell precursors (Fig. 1A, CD20 + CD10+, gated in blue) in terms of the intensity of CD20 and CD10 expression. While circulating immature B-cells were detected as early as Day 3 post CAR-T, at this early time point, the number of B-cells was typically very low (Patient 3 with 0.003% B-cells and Patient 7 with 0.06% B-cells, Table 2). Patients 5, 8 and 9 (Table 2), showed delayed B cell recovery; circulating B cells were not detected until 2–3 months’ post infusion, at which time the vast majority were CD10(+). Interestingly, CD10 gradually down-regulated over time. FCM of post-infusion PB showed slightly decreased CD10 expression at Day 28 (Fig. 1E), compared to Day 13 and Day 15 (Fig. 1C and D). The circulating immature B-cells also expressed CD20 and bright CD45 (at an intensity characteristic of mature lymphocytes), bright CD9, moderate CD22, moderate CD38, dim CD58, polyclonal light chain sIg and were negative for CD34 (Fig. 2C).

Table 2.

Distribution of B-cell Populations in Bone Marrow and Peripheral Blood Status Post CD19CAR-T.

		Bone Marrow Aspirate post-CAR-T					Peripheral blood post-CAR-T
		CD19 (+) B-cells		Marrow B-cell populations (% of B-cells)			CD19 (+) B-cells		Circulating CD10 + CD20+B-cells (% of B-cells)
Patient	Days post CAR-T	% cells	% MNC	CD10 + CD20−/low	CD10 + CD20+	CD10−CD20+	% cells	% MNC	CD10 + CD20+	B-ALL
1	60	1.4	11.3	42	55	<1	1.3	4.6	70	(−)
	75	7.8	36.8	77	18	<1	3.4	5.7	39	(−)
2	13			Not performed			1.8	5.3	>99	(−)
	15			Not performed			1.7	6.4	94	(−)
	28	10	40	>99	<1	<1	5	15	78	(−)
3	aPre CAR-T/28	20	46	86	12	<1	0.25	1.5	71	(−)b
	3			Not performed			0.003	0.04	12	(−)
	28	0.07	0.5	87	<1	<1		cNo B-cells detected		(−)
4	28	2.4	17.5	>99	<1	<1	0.01	0.04	58	(−)
5	28	0.7	5	>99	<1	<1	0	0	0	(−)
	60	4.7	13.5	58	34	<1	1	1.9	>99	(−)
	7d			Not performed			3.4	11.3	62	eMRD (+)
	19d	9.3	39	B-ALL (30% of MNC)		<1	1.6	7.6	75	fB-ALL (+)
6	35	20	58	79	15	<1	3.9	7.2	>99	(−)
7	3			Not performed			0.06	0.57	23	(−)
	28	5.1	35	>99	<1	<1	0.07	0.25	38	(−)
	35			Not performed			0.39	1.4	93	(−)
8	90g	4.2	18	56	40	<1	5.7	8.5	95	(−)
9	75h	0.29	1.1	43	56	<1	0.17	0.61	93	(−)
Median		4.9	27	79	15	<1	1.2	3.3	73
95% CI		3.9–10.4	22–32	62–91	17–25	<1	0.8–2.5	2.2–6.4	52–83

Abbreviations: CAR-T = chimeric antigen receptor T-cell, MNC = mononuclear cells, B-ALL = B-lymphoblastic leukemia, MRD = minimal residual disease.

^aNotes: Day 28 post chemotherapy, prior to CAR-T therapy;

^bB-ALL MRD detected in BM only (0.3% of MNC) in Patient 3.

^cThis time-point excluded from calculation of median and 95% CI, due to lack of identifiable CD19 positive B-cells.

^ddays post second CAR-T cell infusion (nearly 4 months after initial CAR T cell infusion).

^eB-ALL MRD detected (0.035% of MNC) in Patient 7.

^fB-ALL detected (1.7% of MNC) in Patient 7.

^gPatient 8 peripheral blood on Day 5, 14, 28, and 35 post CAR T cell infusion showed only rare CD19 positive B cells, at levels too low to further characterize.

^hPatient 9 peripheral blood on Day 9, 13, and 28 post CAR T cell infusion showed only rare CD19 positive B cells, at levels too low to further characterize.

Fig. 1.

Circulating immature (CD10+) B cells; CD19 + CD20+ B cells (blue); B-cell precursors/hematogones (green). A, Normal B-cell precursors in marrow. B, Rare circulating CD10+ B cells in concurrent blood. C-E, Circulating immature (CD10+) B-cells at various time-points post CAR-T (Patient 2).

Fig. 2.

Representative FCM of B-ALL (Patient 1) with circulating immature (CD10+) B cells post first CD19 CAR-T cell infusion; B-ALL blasts (red), CD19 + CD20+ B cells (blue), B-cell precursors/hematogones (green). (A) Pre-treatment B-ALL immunophenotype. (B) MRD negative marrow, with abundant B-cell precursors with progressive maturation. (C) Circulating immature (CD10+) B-cells in concurrent peripheral blood.

Distinguishing Circulating Immature B-Cells from B-ALL

Initially, the prominent population of circulating CD10(+) B-cells raised concern for involvement by B-ALL; however, they were distinguishable by FCM from the patient’s previously characterized B-ALL, even if disease was originally CD10(+), as illustrated by Patient 5 (Fig. 3, Table 1). Prior to CAR-T, this patient’s B-ALL expressed CD19, CD34 (partial), bright CD10, and bright CD20, but was negative for CD58 and sIg (Fig. 3A). Sixty days after the first CAR-T cell infusion, no evidence of B-ALL was seen in BM or PB. Instead, B-cell precursors with evidence of progressive maturation were abundant in BM (Fig. 3B) and all circulating immature B-cells demonstrated polyclonal surface light chain expression (Fig. 3C). Due to residual B-ALL detected in CSF (data not shown), the patient received a second CAR-T infusion, with relapse 19 days later (Fig. 3D). At that time, blasts remained CD10(+), partial CD34(+), and negative for CD58 and sIg, with slightly downregulated CD20 and CD19 in both BM and PB (Fig. 3D and E). Circulating immature B cells were still detectable at that time point, and remained CD20(+), with polyclonal sIg. A slight decrease in the intensity of CD10 expression was observed in PB at Day 19 post second CAR-T infusion (Fig. 3E) when compared to PB at Day 60 post initial CAR-T infusion (Fig. 3C), suggestive of continued B-cell maturation.

Fig. 3.

Circulating immature (CD101) B cells in a B-ALL patient with relapse after second CAR-T cell infusion (Patient 5); B-ALL blasts (red), CD19 + CD20+ B cells (blue), and B-cell precursors/hematogones (green). (A) Residual B-ALL in bone marrow prior to CAR-T; expressing CD19, CD34 (partial), CD10 (bright), CD20 (bright) and CD45 (dim to negative); negative for CD58 and light chain sIg. (B and C) Bone marrow and peripheral blood FCM, Day 60 post first CAR-T infusion; MRD negative; B-cell precursors with progressive maturation (B); circulating immature (CD10+) B cells (C). (D and E) B-ALL relapse, Day 19 post second CAR-T infusion, with persistence of circulating immature (CD10+) B-cells; immunophenotypic shift in blasts (D) with downregulation of CD19 and CD20 compared to pre-treatment (A) but retaining expression of CD34, bright CD10; persistence of circulating immature (CD10+) B-cells (E) with slight downregulation of CD10 compared to 60 days post first CAR-T infusion (C), indicative of maturation, and distinguishable from circulating B-ALL blasts.

Circulating immature B-cells expressed polyclonal sIg, although we did identify skewing of kappa surface light chain expression in 2 patients (Patient 3 and 7). Prior to CAR-T, the B-cells for Patient 3 expressed CD10, with an appropriate kappa: lambda ratio (kappa: 80% vs.lambda: 19%). Three days post CAR-T, the B-cells showed a marked kappa sIg predominance (kappa 93%: lambda 6%), with very low numbers of circulating B-cells. Patient 7 exhibited circulating immature B-cells with kappa sIg skewing on Day 28 post CAR-T; however, by Day 35, B-cells had increased and skewing was less pronounced. In both instances, the transient skewing was associated with marked B-cell lymphopenia. The number of circulating total B-cells and circulating immature B-cells relative to days after CAR-T therapy is shows in Figures 4 and 5, respectively. When B-cell lymphopenia was observed, it was typically within the first 28 days of CAR-T therapy.

Fig. 4.

The relationship between the amount of total peripheral blood B-cells, as % mononuclear cells (MNC) and the number of days after CAR-T therapy is shown. Each patient timepoint is indicated by a colored dot/line. Patient 5 received two CAR-T infusions; the first infusion is indicated as 5 A, and the second infusion is indicated as 5B.

Fig. 5.

The relationship between the amount of circulating immature (CD10+) B-cells (as % of the total circulating B-cells) and the number of days after CAR-T therapy is shown. Each patient timepoint is indicated by a colored dot/line. Patient 5 received two CAR-T infusions; the first infusion is indicated as 5 A, and the second infusion is indicated as 5B.

Circulating Immature B Cells Independent of CAR-T Therapy

Surprisingly, in Patient 3, circulating immature B-cells comprised 71% of B-cells (Table 2) 28 days post chemotherapy, and prior to CAR-T. The PB was MRD negative; the concurrent BM showed a small amount of B-ALL (0.3% of MNC) in an abundant background of regenerating B-cell precursors. This finding suggests that circulating immature B-cells, which we frequently observed in the context of CAR-T, are not restricted solely to this therapy.

B-Cells in BM

We examined concurrent BM aspirates by FCM at timepoints where circulating immature B-cells were identified. Typically, BM aspirates were analyzed Day 28 post CAR-T infusion and monthly afterwards. By FCM, all 9 BM samples were MRD negative for B-ALL at Day 28. We observed relapsed disease in 1 patient after second CD19CAR-T infusion (Fig. 3D). With one exception, all patients had more BM CD19(+) B cells post CAR T cell infusion than in the corresponding PB (median 4.9% in marrow vs.1.2% in blood, P = 0.001, Table 2). For all 9 cases, FCM detected abundant BM B-cell precursors at time points where circulating immature B-cells were concurrently identified. The majority of B-cell precursors were CD10(+)/CD20 negative to dim (Fig. 2B, gated in green) with early phase maturation (median 79% of B cells in marrow, Table 2) corresponding to the late Pre-B-I (stage 4) and early Pre-B-II (stage 5) designation. Frequently, mature B-cells (i.e., CD10 (−) CD20(+)) were either rare or not detected in the BM. Interestingly, the circulating immature B-cells were also frequently detected in BM, albeit at lower levels (median 15% of BM B-cells vs.73% of PB B-cells). Figure 6 demonstrated the mean percentage of the various B-cell precursor compartments, and their relationship to time after CAR-T therapy. At Day 28, the highest proportion of B-cell precursors was comprised of CD10(+)/CD20(−) cells corresponding with Pre B-I stages; however, 35 days (and greater) after CAR-T therapy a larger proportion of B-cell precursors that were CD10(+)/CD20(+) (progressing to late Pre-BII stage) was observed.

Fig. 6.

The relationship between various bone marrow B-cell precursor compartments and number of days after CAR-T infusion is demonstrated. B-cells precursors expressing CD10(+)CD20(−)/low (Pre-B-I cells) are shown in blue. B-cells precursors expressing CD10(+)CD20(+ (late Pre-B-II cells) are shown in orange. Mature B-cells that are CD10(−)CD20(+) are shown in grey.

DISCUSSION

We observed significant circulating CD10(+) B-cells in B-ALL patients post CAR-T; immunophenotypically, these are consistent with regenerating/non-neoplastic immature B-cells. Since B-ALL is frequently CD10(+), the presence of these circulating immature B cells was surprising, and initially posed a diagnostic challenge. The appearance and timing of these cells was variable in our series, appearing as early as 3 days, and as late as 3 months after CAR-T infusion. Irrespective of timing, and despite the frequent lymphopenia seen in these patients, the proportion of circulating immature B-cells was consistently high, generally >50% of B-cells. In fact, over half of these patients showed 93% to >99% of circulating B-cells expressing CD10, a strikingly unusual finding. There was a strong association between the presence of circulating immature B-cells, and marrow FCM consistent with an early regenerative B-cell picture, one that was abundant in B-cell precursors, and lacking in mature B-cells.

Circulating immature B-cells have been previously reported in various clinical settings and disease states, including in normal individuals. Rare circulating immature B-cells were detected in 66 of 102 (65%) normal adults at 0.06% (median) of total cells and <5% of B cells (18), with no apparent age correlation (19). This is in contrast to B-cell precursors in bone marrow, which are higher in number in children compared to adults (20). Circulating immature B-cells are more abundant in neonates (median 18% of B-cells) (21) and may comprise 10–15% of B cells in immunocompromised and autoimmune-disease patients, as well as bone marrow transplant recipients (22,23). The highest proportion of circulating immature B-cells are reported in the post-Rituximab setting (median 51% of B-cells) (24–27). A significant proportion of circulating immature B-cells was detected in B-ALL patients post therapy (28), at 10–30% of B-cells, occurring within 2 years of therapy. We noted that in Patient 3, at 28 days post chemotherapy and prior to CAR-T therapy, a significant proportion of circulating B-cells were immature (CD10+); this is consistent with literature indicating that this phenomena has previously been observed prior to the advent of CAR-T therapy.

B-cell differentiation in bone marrow is well described, and the normal patterns of hematogone maturation have been elegantly demonstrated by FCM (17,29). B-cell differentiation may be subdivided into 8 stages based on expression of CD19, CD22, CD34, CD10, CD20, CD38, TdT, sIg and cytoplasmic Ig, with stages 1–2 corresponding to Pro-B cells, stages 3–6 corresponding to Pre-B cells, stage 7 corresponding to Immature B-cells, and Mature B-cells in stage 8 (17). The B-cell precursors in healthy children are comprised pre-dominantly of Pre-B-II, Stage 5–6 cells (17), and the number of B-cell precursors in children is significantly higher (median 7.26%) than in healthy adults (median 0.98%) (20), and also holds true in both children (24.6%) and adults (6.3%) with various disease states (29). The patterns of B-cell regeneration in the bone marrow of ALL patients can also reflect the type and timing of chemotherapy that was received. More intensive treatment resulted in a greater proportion of earlier (Pre-B-1 cells, Stage 3–4, CD10+/TdT+) precursors in the hematogone compartment (28,30), and the greatest expansion of B-cell precursors occurs with the cessation of therapy (30). The timing of this expansion is difficult to precisely pinpoint, but marrow B-cell precursors appear to peak in the marrow within approximately 6–12 weeks post cessation of therapy, and the expansion may be prolonged, persisting for up to 2 years (28,30). Recovery of both marrow B-cell precursors and detection of circulating immature B-cells, in pediatric patients with ALL post treatment, was described by Van Wering et al. (28). They showed that at 3 months post therapy, percentages of CD10(+) B-cell precursors (Pre-B I/II, Stages 3–6) increased from 2% to approximately 30% (mean), with a concurrent increase in circulating immature cells expressing CD10 (10–30% of B-cells) corresponding to an increased recovery of total B-cells in blood.

The striking feature in our series of pediatric B-ALL patients post CAR-T, is that the proportion of circulating immature B-cells is far greater (median 73%, 52–83 95% CI) than previously reported by van Wering et al. and the timing of this phenomena appeared earlier, within days after CAR-T therapy. One patient showed circulating immature B-cells 3 days after therapy. A third of our patients showed circulating immature B-cells within 15 days after CAR-T therapy, with most of the patients demonstrating circulating immature B-cells by Day 35. Although recovery of B-cell precursors is age dependent, the observations from this series were unlikely attributable to age alone, as our patient were all under the age of 25, and the study by van Wering et al. included patients aged 16 and below. The majority of patients in our series also showed progressive increase in B-cell recovery with increased time after CAR-T infusion which is compatible with the recovery of B-cells noted in other studies after cessation of therapy (28,30). Of interest, we noted a decrease in peripheral blood B-cells in Patient 5 at Day 19 after second CAR-T infusion (Fig. 4, patient 5B), despite persistence of a substantial proportion of circulating immature B-cells (Fig. 5, patient 5B). This Day 19 timepoint also corresponded to the recurrence of the patient’s B-ALL (Table 2). This is consistent with literature showing that the presence of neoplastic cells, including (but not limited to) ALL, may suppress recovery of marrow B-cell precursors (20,29). One limitation to our series is that it is small; larger studies may be helpful for fully delineating in greater detail the phenomena we have uncovered in the context of CAR-T therapy.

The presence of a homogeneous population of CD10+ circulating B-cells was striking, and initially concerning for involvement by B-ALL. However, careful FCM examination distinguished this population from MRD based on the following features: (1) Comparison to diagnostic IP: although these B-cells expressed CD19, CD10, moderate CD20, bright CD9, moderate CD22, and moderate CD38 that may have overlapped with the original B-ALL IP, they were CD45 bright(+) and CD34 (−) in all 9 cases. The intensity of CD45 was consistent with that of mature lymphocytes, and was a helpful contrast to the dim/dim-to-negative CD45 commonly seen in most B-ALL. Additionally, the circulating immature B-cells had a consistent and uniformly recognizable IP across all patients; in contrast, the blast IP is often different and unique to each B-ALL patient. Furthermore, circulating immature B-cells are recognizable in the presence of CD10(+) B-ALL relapse (Fig. 3E). (2) Polyclonal sIg expression: circulating immature B-cells in all 9 cases clearly showed polyclonal surface light chain expression, in sharp contrast to B-ALL cells that were uniformly negative for sIg in our series. Transient kappa surface light chain skewing was observed in rare cases post CAR-T, but was associated with a marked B-cell lymphopenia and appeared to resolve with time and progressive B-cell recovery. One caveat – B-ALL can occasionally express monoclonal light chain sIg; however, this is quite rare. 3) Follow-up: Examination of sequential FCM post CAR-T may reveal IP features of the circulating immature B cells (downregulation of CD10), consistent with progression to mature B-cells. Correlation with BM showing FCM evidence of progressive B-cell maturation is also a reassuring finding.

Antigenic shifts may occur in B-ALL after therapy. A comparison of B-ALL diagnostic IP vs. relapse IP showed that CD19 is most stable, while CD34 and CD20 differ significantly in nearly half of the cases (31). Interestingly, despite these changes, end-of-induction MRD assessment is still reliable, due to our ability to identify a IP-aberrant population (whether identical, or slightly different from the B-ALL diagnostic IP) from normal B-cell precursors with a characteristic FCM pattern of maturation. In our study, we observed a slight down-regulation of CD19 and CD20 at relapse after the second CAR-T cell infusion in Patient 5. Loss of CD19 is seen in 10–20% of patients post CAR-T, is associated with relapse, and is due to mutation and/or alternative splicing of the CD19 gene (32). Nevertheless, the blasts in our cases retained bright CD10 expression, which was consistent with the diagnostic IP. Therefore, for MRD detection, it is imperative to assess all abnormal patterns of antigen expression and not use “fixed gates” on aberrant blasts from the initial/diagnostic specimen; otherwise, MRD may be missed. Utilization of CD22 and CD24, according the method of Cherian et al. is a useful approach for MRD detection of CD19 negative/silent disease (15). Other markers such as CD86, CD49f, and CD304 may be potentially useful for MRD detection (33–37). Patients in our series received peripheral blood FCM monitoring weekly post CAR-T infusion; however, weekly staging is not routine practice for typical management of B-ALL. Currently, assessment of kappa/lambda sIg is not included in the Children’s Oncology Group recommended approach for B-ALL MRD detection. If CD19 CAR-T therapy becomes more widespread, and weekly FCM monitoring of blood becomes routine, inclusion of light chain sIg in the FCM panel may help to resolve the issue of any CD10(+) B-cells that may circulate post infusion.

Circulating immature B cells that are CD10(+), CD38 bright(+), CD20(+), CD45(+), and CD27(−) reportedly correspond to transitional/immature B cells; they can be further subdivided into early, intermediate and pre-naïve transitional cells (T1, T2, T3, respectively) and characterized by their expression patterns and intensity of CD38, CD27, CD21, CD23, CD44, CD81, sIg heavy chain (e.g., IgM, IgD), CD24, and CD5 (22,38–40). Recently, a similar B cell subset was reported to have regulatory functions, different from traditional transitional/immature B cells (41,42). In our series, the circulating CD10(+) B cells expressed CD38 and CD24 (CD27 not performed). Additionally, in all corresponding marrow aspirates, B-cell precursors were abundant, with a subset expressing CD10 and CD20, as expected for regenerating hematogones. We also observed skewing of sIg light chain expression towards kappa sIg in 2 of the patients in our series, that eventually resolved with time and increased B-cell recovery. A higher kappa/lambda ratio has been demonstrated in circulating immature B-cells, when compared to naïve and memory B-cells, the subsets that comprise the vast majority of circulating B-cells (40). Therefore, we infer that the circulating CD10(+) B cells in our series are immunophenotypically compatible with immature/transitional B-cells. Additional investigation to include some of the immunophenotypic markers described above may be helpful in future studies for further characterization of this B-cell population.

Circulating immature B-cells exhibit heterogeneity, and likely include B-cells exiting the marrow prior to fully reaching the naïve mature B-cell stage. It is proposed that appropriate development of circulating transitional B-cells is bruton tyrosine kinase dependent, and further evidence suggests the presence of an immune checkpoint to reduce autoreactivity within the transitional B-cell stage (40,43). A significant amount of newly circulating B-cells (approximately 40%) are potentially auto-reactive (44), but normally, reduction in autoreactive transitional cells that does occur. Conversely, this may explain why increased transitional B-cells are so frequently observed in a variety of autoimmune conditions (e.g., systemic lupus erythematosus, common variable immunodeficiency), and immunodeficiency states (X-linked lymphoproliferative disease) (22). Furthermore, increased transitional B-cells are also associated with B-cell regeneration states, that may occur after B-cell depleting interventions. This includes cessation of chemotherapy, targeted B-cell therapy (e.g., rituximab), post-transplant setting, or essentially any intervention where regenerating B-cells are re-constituting the marrow, lymphoid tissue, and blood compartments. Intensity and duration of chemotherapy can affect the patterns of marrow B-cell maturation (30), and increased circulating transitional B-cells may reflect a florid B-cell regeneration in the marrow after therapy (28). We surmise that as newer targeted therapies and additional CAR-T cell constructs continue to be developed, this phenomenon will continue to be observed in the future.

In summary, we highlight a series of B-ALL patients developing CD10(+) circulating immature B cells post CAR-T therapy. Detailed FCM of blood and marrow distinguishes this non-neoplastic population from B-ALL MRD. Our findings are relevant to CD19CAR-T therapy with a specific CAR construct utilizing a single-chain variable fragment plus TCR zeta and CD28 signaling. This construct exhibits limited persistence, with disappearance of the CD19-CAR T cells in all patients by Day 53, and recovery of normal B cells (7). It is possible that our findings may not be replicated in situations where alternative CAR constructs with longer persistence are utilized.