Chronic Lymphocytic Leukemia (CLL) is characterized by the accumulation of monoclonal B lymphocytes expressing CD5, CD23, and low surface immunoglobulin in blood, bone marrow, and lymphoid tissues (1, 2). Analyses of many CLL clones demonstrated that the IG gene rearrangements encoding the CLL B cell receptor (BCR) exhibit a striking skewed use of IGHV genes resulting in an IG repertoire different from that of normal B lymphocytes (3–5). Moreover, despite the enormous diversity potentially generated by the recombination of IGHV-IGHD-IGHJ genes, up to 40% of the CLL clones (6, 7) exhibit highly similar stereotyped BCR, which has led to the categorization of the CLL clones with stereotyped BCR into subsets based on their similarities. Although several hundred CLL subsets have been identified, those most frequently encountered, defined as “major subsets,” are limited in numbers. Stereotyped BCRs are determined based on the VH CDR3 features of at least 50% of amino acids identity and 70% of amino acid motif similarities, identical VH CDR3 length and location of a shared pattern(s) among sequences of the same group (6, 8), and the use of IGHV genes belonging to the same phylogenetic clan. IGHV clans are IGHV family genes with structural similarities (9). Conceptually, the stereotyped gene rearrangements should be considered part of the public IG repertoire because different individuals share them (10). In addition, CLL clones of the same stereotyped subsets show IG light chain restrictions (e.g., #1, #2, #4, #6, #8, #64b, and #99) (6, 11–13) and IGKV-IGKJ/IGLV-IGLJ gene rearrangements presenting stereotypy features similar to those of IGHV rearrangements with limitations in IG light chain gene usage and VL CDR3 composition.

The above observations support the notion of a role for BCR stimulation in CLL ontogeny (14–16); moreover, the results of therapies with inhibitors of BCR-associated kinases suggest that stimulation via BCR may be critical for the survival/proliferation of CLL cells in full-blown leukemia (17).

Previous studies have identified IGHV-IGHD-IGHJ rearrangements sharing features with that characteristic of CLL subsets in splenic and circulating B cells from normal subjects (18–20). These rearrangements, from now onward defined as CLL-like stereotyped-IG or CLS-IG, can be observed in different B cell subpopulations, even though they accumulate in the CD5⁺ B-cells (20, 21).

This study used high-throughput sequencing technology on peripheral blood B cell subsets to elucidate CLS-IG’s features and cellular distribution. For this purpose, the cells, separated into defined subsets, were also fractionated according to CD5 or IG light chain expression. The data obtained provide a new perspective for interpreting the origin of CLL cell repertoire and possibly for disease ontogeny.

The identification of CLL clones’ stereotypical IG receptors indicates that different individuals share immunoglobulin rearrangements with these specific features. This aspect reminds the existence of identical CDR3 aa sequences between individuals, which defines the so-called public repertoire (10, 29). More recently, the definition of public repertoire has been made less stringent by including rearrangements with a high degree of similarity (30). In this respect, the public repertoire could enclose stereotypic rearrangements recognized in CLL clones.

The scope of this study was that of investigating whether, in healthy donors, CLL IG stereotyped sequences were present in the IG repertoire of any of the circulating B-subsets and to which extent. The donors analyzed did not have evidence of lymphocytosis, thus excluding the possibility of introducing biases due to some type of preleukemic stage, which may result in B-cell repertoire alterations (31). Blood samples were collected from donors with a median age of 58.8 (range 56-67), which is in the range of possible identification of clonotypes ascribable to CLL clones that became clinically evident later. Indeed, it has been reported that clonotypes attributable to the leukemic clone can be identified in the PB up to 16 years before CLL diagnosis (32). It is, however, unknown whether phenotypic changes were present in the PB B lymphocytes at this early stage. The age choice appears relevant for comparison with CLL, given that changes in B cell repertoire within single B-cell subsets may occur with advancing age and may influence the cell population in which leukemogenesis occurs. In addition, it has been reported that changes in the representation of certain CLS-IG can be observed in aging individuals (19).

Although CLL stereotyped receptors have been reported in healthy donors (18, 19, 33–35), our approach provides more details on their phenotypic features, considering, for instance, the expression of IG light chains and CD5 expression within B-subsets. The identification and characterization of CLL stereotyped IG in healthy donors could provide information on the cell(s) of origin of CLL. Attempts to identify a specific B-subset as a compartment of origin for CLL have been unsatisfying. Here we attempt to lay the ground for a better understanding of the CLL cell of origin.

Consistent with previous reports (20, 21), our data show a significantly higher presence of CLS-IG in bulk CD5⁺ compared to bulk CD5^- B cells. Nevertheless, analysis of U- and M-IGHV gene rearrangements within CD5⁺ and CD5^- B cells highlighted no enrichment of CLS-IG in CD5⁺ compared to CD5^- B cells within each mutational category ( Figure 2B ). Instead, CLS-IGs were enriched in U-IGHV genes regardless of being expressed by CD5⁺ or CD5^- B cells.

CLS-IG representation was also investigated in B-subset and their respective CD5⁺ and CD5^- fractions. The analysis showed a higher CLS-IG percentage in TR and N than in MZ and SM B-subset ( Figure 3A ), a finding in line with a previous study from our group on splenic CLS-IG carrying the IGHV-1 family genes (20). Since TR and N cells utilize almost exclusively U-IGHV rearrangements ( Figure 3C ), the higher representation of CLS-IG in these cell fractions is likely related to the predominance of U-IGHV rearrangements. The analysis of CD5⁺ and CD5^- cells from each B-subset separately showed no dominant presence of CLS-IG in any of the cell fractions studied ( Figure 3B ). However, MZ shows an evident trend where CD5⁺ have more CLS-IG than CD5^-. It must be noticed that MZ presents a considerable amount of U-IGs concentrated in the CD5+ fraction ( Figure 3C ). It must be mentioned that because CD5⁺ B cells are scarce in predominantly mutated B cells subsets, a limited number of IG rearrangements were obtained in some CD5⁺ B-cell fractions (i.e., MO, SM, DN), impairing the possibility of statistical analysis (see Supplementary Table S2 for details). Altogether, our data do not present evidence of a phenotypically distinct B-subset with a significant accumulation of CLS-IG. Their predominance in CD5⁺ N and TR B cells depends upon the higher presence of U-IG rearrangements in these B-cell fractions.

The relative distribution of CLL subsets in normal B cells did not follow that typically reported for CLL ( Figure 4B ). For example, CLL subsets #4 and #8, relatively frequent among the CLL major subsets, were rarely (or never) identified. A preferential representation of individual CLS-IG in the IGκ or IGλ expressing B-subset was not observed. In contrast, specific CLL subsets (e.g., #1, #2, #4, #6, #8, #64b, and #99) show, in CLL, marked IG-light chain use restrictions (6, 12, 36). It might be worth mentioning that CLS-IG #2 showed a trend indicating a higher representation in IGκ⁺ normal B-cells. This would be compatible with the observation that, in some instances, CLL clones belonging to subset #2 show functional rearrangements of the IGκ light chain inactivated with a Kappa Deleting Element (37). Thus IG-receptor editing may operate in this apparently large proportion of CLS-IG subset #2/IGκ⁺ B cells. These observations indicate no major structural constraints contributing to certain IGH-IG light chain pairing in the CLS-IG of a normal B-cell repertoire. This is in keeping with what has been reported about the possibility of CLL subset #2 IGHV rearrangement to pair with “non-native” (e.g., other than IGLV3.21 encoded IGλ light chain) IG light chains (38).

The IGHV genes used by CLS-IG rearrangements within each CLL subset showed a heterogeneous pattern. Their overall utilization was not always as restricted as CLL stereotyped IG. For instance, the IGHV1-69 gene was virtually absent in CLS-IG from CLL subset #1, even though this gene is one of the IGHV1 genes highly represented in control IGs. CLS-IG in CLL subsets #1 and #3 exemplify the absence of IGHV selection, whereas CLS-IG in CLL subset #14 acts closely to what is observed in CLL with a highly prevalent representation of the IGHV4-4 gene. This indicates that specific VH CDR3 sequences may have a non-random association with IGHV genes within a VH clan, possibly related to restrictions during IGHV-IGHD-IGHJ rearrangements and/or positive and negative selection in ontogenesis or the course of early antigenic challenges. The absence of IGκ and IGλ restriction for each of these CLS-IG subsets points out, at this stage, negligible participation of the IG light chains. Altogether, the leukemogenic process likely involves a further selection of IGHV genes and the IG light chains in a CLL stereotype-specific manner.

The mutational status of CLS-IG deserves particular comment. When the CLS-IG analysis was extended by removing the IGHV mutational status as a prerequisite for classification (see Methods), the representation of CLS-IG mainly followed the mutational status characterizing the original CLL clones. For instance, subset #1 (always unmutated in CLL) was identified predominantly in the U-IG repertoire of normal donors and the majority of subset #14 CLS-IG (mutated in CLL) were recognized in the M-IG repertoire. Likewise, subset #2 CLS-IG were found in the U and M CLS-IG repertoire, as observed in the CLL cohorts. The observation that the mutational status of CLS-IG parallels that of CLL IG stereotypes suggests that specific IG rearrangements could influence clonal function, e.g., by limiting the generation of a post-germinal center progeny (CLL subset #1) or by determining an accumulation of memory B cells (CLL subset #14).

Thus, single CLS-IG in recirculating B cell has only marginally superimposable features compared to those encountered in CLL clones and each one appears to have its characteristics. For instance, subsets #4 and #8 are substantially absent in the CLG-IG repertoire; subset #1 CLS-IG shows the utilization of IGHV genes closer to CLL stereotypes, whereas that of CLS-IG, subset #3 is more random. It can be presumed that the trajectory determining the emergence of CLL clones is very heterogeneous.

The above data demonstrate that CLS-IG detected in peripheral B cells from donors with no evidence of peripheral lymphocytosis have different features than those typically identified in leukemic clones, suggesting the shaping of CLL BCR repertoire and the emergence of the leukemic clones is dictated by numerous selecting factors. A recent study (32) showed that skewing of the B-cell repertoire is observable in some clusters before the clinical presentation of CLL. Thus, subjects in a pre-leukemic phase or predisposed to developing CLL are likely to have a different condition than the donors analyzed here. The observations that CLS-IGs are not enriched in PB derived B cells with a defined phenotype can be interpreted in different ways: 1) CLL cells originate from non-circulating B cells residing in solid lymphoid tissue, 2) CLL cells originate from B cells with a different phenotype than the ones explored in this study, 3) CLL cells originate from B cells without a defined immunophenotype. The above interpretations may not be mutually exclusive. In addition, it is possible that the leukemogenic process can be accompanied by immunophenotypic changes that encompass an elevated expression of CD5 typically observed in CLL and MBL cells and presumably the pre-monoclonal B-cell lymphocytosis described by Kolijn, P. et al. (32)

In this context, B-cell immune-genotype appears to be a relevant factor in the quest to identify the CLL clones’ cell of origin, adding additional elements useful for understanding CLL emergence routes.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/genbank/, PRJNA807871.

DB designed the study, designed the library preparation protocol, analyzed data and wrote the manuscript. MoC performed FACS Sorting, prepared the library, sequenced and analyzed the data and wrote the manuscript. DR designed and performed FACS Sorting. SM and RM processed biological samples and prepared the library. GU and VA collected samples. LA, SV, and AM performed data and statistical analysis. AN, FM, and GC supervised research. MaC and FG prepared the figures. MF and FF designed the study and wrote the paper. All authors read and approved the final manuscript.

This work was supported by: Associazione Italiana Ricerca sul Cancro (AIRC) ID.15426 (to FF); AIRC and Fondazione CaRiCal co-financed Multi-Unit Regional Grant 2014 n.16695 (to FM); Italian Ministry of Health 5 × 1000 funds [2014 ([to GC), 2015 (to FF), 2016 (to FF and GC), Ricerca Corrente 2016 (to FF and GC); Gilead fellowship program 2016 (MC) and 2017 (GC); Associazione Italiana Ricerca sul Cancro (AIRC) Grant 5 x 1000 n. 9980 (to FM and AN); AIRC IG 24365 to AN; Associazione Italiana Leucemie, Cosenza (to FF). The primers for the determination of IGHV repertoire were kindly provided by TIB Molbiol srl (Genoa, Italy). AM received funding from the European Union’s Horizon 2020 Research and Innovation Programme under the Marie Skłodowska-Curie grant agreement No. 101023721.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.