Evidence of shared epitopic reactivity among independent B-cell clones in chronic lymphocytic leukemia patients.

Chronic lymphocytic leukemia (CLL) is a neoplastic disease of mature B cells that express a functional immunoglobulin (Ig) B-cell receptor (Ig-BCR) on the cell surface.[1] The Ig-BCR includes the binding site (idiotype) for the epitope of cognate antigen, which results from the stochastic and productive Ig variable genes rearrangement and possible somatic hypermutation. Individual B-cell clones differ from one another because of the diverse amino acid sequences within the idiotype of the expressed Ig-BCR,[2] so that the epitopic specificity is shared in a given clonal B-cell population.

Several studies have recently highlighted two peculiar characteristics that distinguish CLL among other mature B-cell malignancies. First, the Ig-BCR repertoire of CLL is biased, as it is restricted toward the preferential usage of certain Ig heavy- (IgVH) and light (IgVκ/λ) -chain variable genes,[3, 4] and unusually similar, stereotyped heavy-chain complementarity determining region 3 (VH CDR3) amino acid sequences.[5] The skewed Ig-BCR could be owing to an Ig-BCR-driven selection mechanism initiated by specific antigens that promote the expansion and possibly the maintenance of the cognate CLL clone.[6] Consistent with this hypothesis, several studies have demonstrated the reactivity of CLL Ig-BCR against foreign antigens, self-antigens, peptides and intrinsic IgVH motif.[7] Second, CLL is not always a monoclonal disorder, as two or multiple CLL clones have been found in 2–5% of CLL patients.[8, 9] Furthermore, the monoclonal B-lymphocytosis precursor state, which precedes the clinically relevant leukemic phase in virtually all CLL patients, sometimes involves multiple B-cell clones.[10]

It is still unknown whether two or more CLL clonotypes within the same patient potentially associate with the same antigenic reactivity, or alternatively arise as stochastic and antigen-independent events, fostered by the accumulation of oncogenic abnormalities in the preleukemic state. To answer this question, here we have characterized the epitope recognition profiles of CLL clonotypes by coupling the genetic analysis of Ig variable genes and the epitopic reactivity at single-cell level.

We isolated single CD5+ B cells from peripheral blood of six newly diagnosed untreated CLL patients, randomly referred to the Medical Oncology Unit—University Magna Graecia of Catanzaro. CLL patients displayed the typical CLL immunophenotype, with no evidence of separate/aberrant B-cell populations (Supplementary Table S1). We determined the VHDJH and VLJL complementary DNA (cDNA) sequence of at least 20 single leukemic cells per patient (Supplementary Table S2). All cDNA sequences showed a productive rearrangement at the heavy- and light-gene loci. In particular, we detected single and distinct VHDJH and VLDL rearrangements in CLL#1, CLL#2 and CLL#3, indicating the presence of a single clonotype (Table 1, Supplementary Table S2). Differently, CLL#4, CLL#5 and CLL#6 exhibited two different VHDJH rearrangements, each of them pairing with a unique and distinct VLJL rearrangement, indicating the coexistence of two clonotypes (Table 1, Supplementary Table S2). The coexistence of two clonotypes in CLL#4, CLL#5 and CLL#6 was confirmed by the heteroduplex analysis and sequencing of VHDJH and VLJL rearrangements amplified from the bulk CLL cells (Supplementary Figure S1). Of note, the Vκ usage of clonotypes CLL5-1, CLL6-1 and CLL6-2 were univocally assigned to Vκ genes of the distal cluster, which is evocative of receptor editing.

Table 1

Characteristics of CLL clonotypes and mimotopes

Patient	Clonotype	mIgCLLa	VHDJH	VLJL	CDR3 lengthb	VH CDR3 IMGT aa sequencec	Freq. (%)d	Stereotypic subsete	Mimotope name	Mimotope aa sequence	KD (nm)f
CLL#1	CLL1	mIgCLL1	V3-11/D3-10/J6	κV1-39(1D-39)/J2	17	AREGLWFGELSYYGMDV	100	NA	pCLL1	CSPAKELGC	34.03
CLL#2	CLL2	mIgCLL2-1	V3-73/D3-3/J6	κV1-33(1D-33)/J1	20	TFDFWSGYYDGYYYYYGMDV	80	NA	pCLL2-1	CNTYSVSLC	10.81
		mIgCLL2-2				TFDFWSGYYDGYYYYYGLDF	20		pCLL2-2	CKSYSVSLC	7.57
CLL#3	CLL3	mIgCLL3-1	V3-23/D3-22/J4	κV1-8(1D-8)J3	17	AKRDYSHRSDYAPLFEY	90	NA	pCLL3-1	CPPQSVTEC	22.14
		mIgCLL3-2				GKRDYSHSSDYAPLFEY	10	NA	pCLL3-2	CDVWHSAYC	5.64
CLL#4-1	CLL4-1	mIgCLL4-1	V4-34/D7-27/J2	κV1-33(1D-33)/J1	17	ARRGTGDPPYWYFDL	77	NA	pCLL4-1	CTTNPADSC	5.50
CLL#4-2	CLL4-2	mIgCLL4-2	V4-4/D6-19/J2	κV2-28(2D-28)J1	21	ARGTVGQQWLEVLDWYFGL	23	NA	pCLL4-2	CVLWWSPIC	3.90
CLL#5-1	CLL5-1	mIgCLL5-1	V4-34/D3-22/J4	κV1D-12/J1	20	ARGGNNDKIVMLLYYFDF	57	NA	pCLL5-1	CFSDDEWWC	7.08
CLL#5-2	CLL5-2	mIgCLL5-2	V4-59/D3-22/J3	κV1-13(1D-13)/J2	18	ARDYDYDTRKSDAFDIW	43	NA	pCLL5-2	CPPFTNYEC	7.32
CLL#6-1	CLL6-1	mIgCLL6-1	V1-46/D3-10/J6	κV3D-15/J1	22	ARDWVATMVRGVIESRPTGMDV	65	NA	pCLL6-1	CNQDWHKKC	57.53
CLL#6-2	CLL6-2	mIgCLL6-2	V4-34/D2-2/J6	κV1D-17/J1	24	ASRFYCSGSSCEAPRYYYYYGLDV	35	16	pCLL6-2	CTTVIPERC	22.67

Name of the CLL-derived monoclonal Igγ1 antibodies expressing the indicated VHDJH and VLJL rearrangements.

Number of amino acids of the VH CDR3.

VH CDR3 amino acid sequence according to the International ImMunoGeneTics information system (http://www.imgt.org).

Percentage of analyzed cells expressing the indicated VH CDR3.

Stereotypic subset according to Agathangelidis et al.[5] NA, not attributable to a currently defined stereotypic subset.

KD values for the mimotope binding to the cognate mIgCLL, as estimated by the Scatchard plot analysis.

The analysis of nucleotide differences among the VH sequences of the same clonotype showed a variable degree of base-pair substitutions (Supplementary Table S3). Despite the intraclonal diversity of the VH region, each clonotype showed either an identical or two quasi-identical VH CDR3 amino acid sequences (Table 1 and Supplementary Table S3). By comparing the VH CDR3 amino acid sequence of clonotypes with a data set of stereotypic Ig-BCRs,[5] CLL6-2 was found to be a member of the major subset 16, whereas other clonotypes did not fall within the stereotypic subsets (Table 1).

Next, we aimed to identify peptide mimetics of the natural epitope (hereafter, referred as mimotopes) that bound to the Ig-BCR of CLL clonotypes. To this end, we produced monoclonal antibodies IgG1 (mIgCLLs) that carried the VHDJH and VLJL rearrangements of CLL clonotypes linked to an IgHCγ1, or IgLCκ/λ segment, as described[11] (Table 1, Supplementary Figure S2). The mIgCLLs were used as baits for screening a constrained (CX7C) random phage-displayed peptide library,[12] which resulted in the isolation of at least 25 phage clones for each mIgCLL bait (Supplementary Table S4). We observed a considerable difference of enzyme-linked immunosorbent assay (ELISA)-binding reactivity among the phage clones isolated through the same mIgCLL (Supplementary Figure S3). As mimotopes are equally distributed among phage particle (five copies per phage), we assumed that the differences in the binding affinity of distinct mimotopes to the same mIgCLL might account for the differences in the ELISA-binding reactivity.

Hence, we synthesized a number of peptides corresponding to the phage-displayed mimotopes that had shown the highest binding affinity toward the cognate mIgCLL (Table 1). Synthetic mimotopes bound in a dose-dependent manner to the cognate mIgCLL, whereas tested negative for control human polyclonal Igs (Supplementary Figure S4), demonstrating that the selected mimotopes maintained the mIgCLL-binding ability out of the phage context. The equilibrium constant KD of synthetic mimotope binding to the cognate mIgCLL was in the nanomolar range, as estimated by the Scatchard plot analysis (Table 1).

All mimotopes bound to the corresponding CLL cells, and not to normal B (sIg+/CD19+) or non-B cells (sIgG−/CD19−) isolated from the peripheral blood of a pool of healthy donors, as shown by the fuorescence-activated cell sorting analysis (Figure 1a). The mimotopes binding to B-CLL cells was competed in a dose-dependent manner by soluble cognate mIgCLLs and not by polyclonal human IgG1 (Supplementary Figure S5). Moreover, mimotopes colocalized with the sIg-BCR on the surface of the cognate B-CLLs, as shown by the confocal microscopy (Figure 1b). Overall, these data confirmed that the mimotope binding to the corresponding B-CLL cells was specifically achieved through direct interaction with the sIg-BCR, thus validating the selected mimotopes as truly representative of the epitopic reactivity of the original CLL cells.

Figure 1

Mimotope-binding profile of CLL. (a) Specificity of mimotope binding to cognate CLL cells by flow cytometric analysis. Non-permeabilized CD5+ cells (1 × 106) from CLL#1 patient (left panel), or sIg+/CD19+ (middle panel) and sIg−/CD19− (right panel) peripheral blood mononuclear cells (PBMCs) (1 × 106) from a pool of healthy donors were incubated with N-biotinylated pCLL1 mimotope (10 pm), or control N-biotinylated scrambled peptide. Peptide binding to CLL cells was revealed by streptavidin-FITC. (b) Colocalization of N-biotinylated pCLL1 mimotope with the surface Ig-BCR of CLL#1 cells. CD5+ cells (1 × 106) from CLL#1 patient (left panel) were incubated with N-biotinylated pCLL1 mimotope (10 pm), stained with strepavidin-FITC (1:100), anti-human-IgM PE (1:100), anti-CD5-APC (1:100) and analyzed by confocal microscopy. Scale bar: 5 μm. (c) Enzyme-linked immunosorbent assay (ELISA)-binding reactivity of selected mimotopes against the mIgCLLs. ‘pCLL' is the labeling for peptide (mimotope) ligand; ‘mIgCLL' is the labeling for the recombinant, CLL-derived Ig. Relative absorbance was calculated as the difference between optical density405 nm and optical density620 nm divided by the absorbance of the blank sample, and reported as shades of green at the bottom. The white square indicates ELISA absorbance value less than twofold relative to blank, indicating no interaction in ELISA.

Then, we performed a systematic characterization of the epitopic reactivity of CLL clonotypes by probing the binding of individual mIgCLLs to all selected mimotopes by ELISA. Each mIgCLL was able to bind not only to its cognate mimotope, but also to distinct mimotopes of others mIgCLLs (Figure 1c). These results highlighted a certain degree of common epitopic reactivity among the mIgCLLs, as previously reported by others.[13, 14]

A first level of common reactivity regarded the mIgCLLs derived from the same patient that invariably shared the epitopic reactivity, regardless they were classified as distinct clonotypes. In particular, the mIgCLLs with quasi-identical VH CDR3, which were derived from patients harboring a unique clonotype (mIgCLL2-1 and mIgCLL2-2 from patient CLL#2; mIgCLL3-1 and mIgCLL3-2 from patient CLL#3), equally bound to the two selected mimotopes (pCLL2-1 and pCLL2-2 for patient CLL#2; pCLL3-1 and pCLL3-2 for patient CLL#3), thus indicating that the small differences in their VH CDR3 did not affect their binding affinity (Figure 1c). The mIgCLLs derived from patients harboring two different clonotypes (mIgCLL4-1 and mIgCLL4-2; mIgCLL5-1 and mIgCLL5-2; mIgCLL6-1 and mIgCLL6-2) also showed common epitopic reactivity, even though with different ELISA reactivity. Thus, the diverse VH CDR3 pattern of distinct CLL clonotypes harbored by the same patient did not prevent the binding to a common set of mimotopes.

A second level of shared epitopic reactivity was observed for some mIgCLLs derived from different patients. In particular, the mIgCLLs derived from patients CLL#1, CLL#2 and CLL#6 recognized the same subset of mimotopes (pCLL1, pCLL2-1, pCLL2-2, pCLL6-1 and pCLL6-2; Figure 1c). The mIgCLL4-2 also shared a similar ELISA-binding profile, except for the binding to pCLL1-2.

The clonotypic mIgCLL1, mIgCLL2, mIgCLL4-2, mIgCLL6-1 and mIgCLL6-2 did not show any obvious similarity in the VH CDR3 primary structure, or a preferential usage of VH and VL regions that would account for the common binding to the same subset of mimotopes. Hence, their shared epitopic reactivity should lie on structural motifs of the whole VH and VL regions, which will require further investigation.

Overall, these data demonstrated that (a) mIgCLLs derived from independent B-cell clones of the same CLL patient shared the epitopic reactivity and (b) a subset of CLLs exhibited some degree of common epitope reactivity.

Our results confirmed recent findings on the appearance of two or more independent B-CLL clones within the same patient.[8, 9, 10] Importantly, we document for the first time that the Ig-BCRs of unrelated clonotypes in the same CLL patient may recognize common epitopes, thus revealing a striking homogeneity, grounded on epitopic reactivity, that was unpredictable from the molecular features of the Ig-BCR, including the CDR3 composition and length. The coexistence within the same patient of two unrelated CLLs clonotypes associated with the same epitopic reactivity reinforces the key role of antigenic selection in CLL pathogenesis and demand a more exhaustive understanding of the biology of this disease based on epitopic reactivity.