Pathobiology of follicular lymphoma

INTRODUCTION — Follicular lymphoma (FL) is the second most common subtype of non-Hodgkin lymphoma. It is defined as a lymphoma of germinal center B cells, and virtually always demonstrates a growth pattern that is partially follicular. (See "Classification of hematopoietic neoplasms".)

The molecular pathogenesis of FL is a complex, multistep process during which a single follicular B cell acquires all of the genetic and epigenetic alterations needed for malignant transformation. The resultant tumor is usually comprised of a mixture of centrocytes (small cleaved germinal center cells) and centroblasts (large noncleaved germinal center cells). Some common steps in this pathway have been elucidated, particularly chromosomal rearrangements involving BCL-2 and certain somatic mutations, some of which are also seen in other non-Hodgkin lymphomas. In the not uncommon instance in which FL evolves into a more aggressive lymphoma, usually diffuse large B cell lymphoma, the molecular lesions found early in the disease course persist and are found along with additional new acquired mutations that contribute to histologic transformation. (See "Histologic transformation of follicular lymphoma".)

This review will discuss the pathobiology of FL in adults. Of note, pediatric follicular lymphoma is a different form of FL that mainly develops in children and adolescents, but is also seen in young adults, and has distinctive clinical and pathologic features. The epidemiology, clinical presentation, pathologic features, diagnosis, treatment, and prognosis of FL are discussed separately. General aspects of the pathobiology of non-Hodgkin lymphoma are also discussed separately. (See "Clinical manifestations, pathologic features, diagnosis, and prognosis of follicular lymphoma" and "Initial treatment of stage I follicular lymphoma" and "Treatment of relapsed or refractory follicular lymphoma" and "Overview of the pathobiology of the non-Hodgkin lymphomas".)

CELL OF ORIGIN — FL is a heterogeneous clinicopathologic entity that has as a common feature an origin from germinal center B cells [1]. The tumor is comprised of variable numbers of small, cleaved cells (centrocytes) and larger, blastoid cells (centroblasts) that morphologically resemble cells found in the light and dark zones of normal germinal centers, respectively. The germinal center ancestry of these cells is principally supported by the identification of somatic mutations in the variable region of the immunoglobulin genes (IgVH), which serves as a marker of germinal center transit [2], the follicular growth pattern of the tumor cells in most cases, and the immunophenotype of the tumor cells, which in most respects closely resembles that of normal germinal center B cells. (See "Overview of the pathobiology of the non-Hodgkin lymphomas", section on 'B cell lymphoma'.)

PATHOGENESIS

Overview — The development of the majority of FL tumors in adults is dependent on the overexpression of B cell leukemia/lymphoma 2 (BCL-2) located on chromosome band 18q21. BCL-2 is an oncogene that blocks programmed cell death (apoptosis). As such, overexpression results in increased cell survival. BCL-2 overexpression in itself is not sufficient for FL development and other genetic lesions or host factors are required (figure 1). (See 'Other genetic lesions' below.)

BCL-2 overexpression is virtually always the result of a translocation that places the BCL-2 gene under the control of immunoglobulin promoter/enhancer elements. Of these, the t(14;18)(q32;q21) is most common and places BCL-2 under the control of the IgH locus, but variant translocations also occur that place BCL-2 under the control of the kappa or lambda immunoglobulin loci instead. (See 'Translocations involving BCL-2' below.)

These aberrations likely represent mistakes that occur during attempted antigen receptor gene diversification in normal B cells. In the bone marrow, pre-B cells rearrange the V(D)J segments of their immunoglobulin (Ig) loci to allow expression of IgM. This process, termed V(D)J recombination, requires the recombination activating genes RAG1 and RAG2. Subsequently, in the germinal center, antigen stimulated B cells undergo class-switch recombination (from IgM to IgG, IgA, or IgE) and somatic hypermutation. Somatic hypermutation, a key component of the humoral immune response, is the process by which point mutations and small deletions or insertions are introduced at high frequency into the variable regions of immunoglobulin heavy chain genes. Both class switch recombination and somatic hypermutation require the enzyme activation-induced cytidine deaminase (AID).

The sequences found immediately adjacent to the points of DNA breakage and rejoining in BCL-2/Ig translocations implicate RAG1/RAG2 in the process [3]. It remains uncertain whether these RAG mediated breaks occur in pre-B cells or during antigen receptor gene "editing" in germinal center B cells. Cells bearing such rearrangements gain a selective advantage when the consequence is overexpression of an oncogene such as BCL-2. AID-mediated class-switching and somatic hypermutation are also error-prone, and it is strongly suspected that additional AID-dependent mutations introduced by these processes in germinal center B cells contribute to the development of FL. (See "Overview of the pathobiology of the non-Hodgkin lymphomas", section on 'Chromosomal translocations'.)

Since FL often shows evidence of ongoing somatic hypermutation, it is possible that AID contributes to acquisition of additional mutations that are involved in histologic transformation of FL as well. (See "Histologic transformation of follicular lymphoma".)

Aberrant BCL-2 expression

BCL-2 action — The B cell leukemia/lymphoma 2 (BCL-2) gene encodes a 26-kDa integral membrane protein localized to the outer mitochondrial membrane and endoplasmic reticulum [4]. BCL-2 has little or no ability to promote cell cycle progression or cell proliferation, but rather raises the cellular apoptotic threshold, thereby inhibiting apoptosis (programmed cell death). Normally in germinal center B cells, BCL-2 expression is restricted to cells that through the process of somatic hypermutation, by chance, acquire Ig with high affinity for antigen; these cells are thought to then emerge from germinal centers as long-surviving memory B cells [4] or to differentiate into plasma cells. This hypothesis is supported by data from mice bearing BCL-2 transgenes that show markedly protracted secondary immune responses and an extended lifetime for memory B cells in the absence of antigen [5,6].

Apoptosis (programmed cell death) is essential for normal embryonic development, maintenance of tissue homeostasis, and development and function of the immune system. In contrast, processes that interfere with normal apoptosis promote cell survival and, potentially, oncogenesis. Apoptosis is regulated by a balance of pro-apoptotic factors and anti-apoptotic factors [7]:

●Pro-apoptotic factors – Two categories of pro-apoptotic factors exist, one consisting of multi-domain proteins (eg, BAX and BAK), and the second consisting of single BH3-domain only proteins (eg, BAD and BID). BAX and BAK are the major initiators of apoptosis. How BH3-domain only proteins promote apoptosis is unsettled; they may contribute by inhibiting anti-apoptotic proteins (see below) or by directly activating BAX. Once activated, BAX and BAK form aggregates in the mitochondrial membrane that promote apoptosis by releasing cytochrome c from the mitochondria into the cytoplasm, leading to the activation of caspase-9 followed by caspase-3, a so-called apoptotic executioner. Active caspase-3 proteolyzes a number of key cellular proteins, including a negative regulatory unit of caspase activated DNase, an endonuclease that cleaves DNA and contributes to apoptosis [8-10].

●Anti-apoptotic factors – BCL-2, BCL-XL and MCL-1 are examples of anti-apoptotic factors. These anti-apoptotic factors restrain the activity of pro-apoptotic proteins, such as BAX and BAK, and prevent the release of mitochondrial proteins, such as cytochrome c, thereby inhibiting apoptosis.

FL demonstrates increased expression of the anti-apoptotic factors (BCL-2, BCL-XL, and MCL1) and very low or undetectable amounts of pro-apoptotic factors such as BAX and BAD [11,12]. This imbalance promotes cell survival.

How BCL-2 translocations contribute to the development of FL is complex and incompletely understood. It has long been known that BCL-2/IgH translocations can be detected at low frequency (ie, one cell in 1000 to one cell in 10,000) in the germinal centers of a high fraction of normal individuals, almost none of whom will go on to develop FL [13,14]. Similarly, highly sensitive methods can detect the t(14;18) at low levels in the peripheral blood of some healthy individuals [15,16]. One study found that these (14;18) translocations are found within a population of atypical germinal center-type B cells that have phenotypic features of FL [17]. These observations clearly indicate that BCL-2/IgH translocations are not sufficient to cause FL, and that other genetic or epigenetic events are rate limiting for its development.

The natural history of persons with low levels of circulating clones resembling FL is not known. In one study, a highly sensitive PCR-based assay was able to identify IgH/BCL-2 fusion genes in a subset of patients up to 15 years before diagnosis of FL [18]. While not all patients with the translocation developed FL, and not all patients who developed FL had the translocation in pre-diagnostic blood samples, samples from those who went on to develop FL were more likely to be positive (56 versus 29 percent).

Translocations involving BCL-2 — In the vast majority of cases, FL is associated with a translocation between the long arm of chromosome 18, the site of the BCL-2 oncogene (18q21), and one of three loci containing immunoglobulin (Ig) genes [19,20]:

●The Ig heavy chain gene on chromosome 14 – Resulting in the t(14;18)(q32;q21) found in approximately 85 percent of FL.

●The kappa light chain gene on chromosome 2 – Resulting in the t(2;18)(p11;q21), which is uncommon but considered a biologic equivalent.

●The lambda light chain gene on chromosome 22 – Resulting in the t(18;22)(q21;q21), which is uncommon but considered a biologic equivalent.

The consequence of these translocations is the presence within the cells of constitutively high levels of BCL-2 protein, resulting from both enhanced transcription and, possibly, more efficient RNA processing (picture 1) [21,22].

BCL-2 negative cases — Less than 10 percent of FLs lack BCL-2 translocations, and in these tumors BCL-2 protein expression is decreased relative to BCL-2 translocation-positive cases or absent altogether [23]. These tumors often fall into one of two morphologic subgroups:

●Predominantly diffuse follicular lymphoma

●Tumors with a follicular growth pattern and grade 3b cytologic features

Predominantly diffuse follicular lymphoma differs from typical FL in showing variable BCL-2 protein expression, strong expression of CD23, frequent deletions involving chromosome 1p36, and mutations in TNFRSF14, a gene encoding a member of the tumor necrosis factor (TNF) family [24,25]. It often presents as a localized mass in the inguinal region and is recognized as a distinct subtype of FL in the World Health Organization (WHO) classification of lymphoid neoplasms [1].

In grade 3b FL sheets of centroblasts grow in a follicular pattern [26]. (See "Clinical manifestations, pathologic features, diagnosis, and prognosis of follicular lymphoma", section on 'Grade'.)

The clinical course of grade 3b FL is generally more similar to that of diffuse large B cell lymphoma (DLBCL) than that of FL grades 1 and 2. One study using cytogenetics demonstrated that this morphologic subgroup includes [27]:

●Tumors with the t(14;18) affecting BCL-2

●Tumors with 3q27 rearrangements affecting BCL-6

●Tumors that lack either of these characteristic cytogenetic abnormalities

The same molecular subdivisions exist in DLBCL, leading to the suggestion that grade 3b FL with the t(14;18) is more similar to FL grades 1 and 2, while the other two cytogenetic subtypes are more closely related to de novo DLBCL [27]. Although this series suggested that BCL-6 rearrangements and BCL-2 rearrangements are mutually exclusive in FL, subsequent studies have identified rare tumors with both rearrangements [28]. In either case, BCL-6 expression may play a role in the pathogenesis of a subset of grade 3b FL. BCL-6 encodes a transcription factor that is essential for normal germinal center B cell development. Expression of BCL-6 in normal germinal center B cells down-regulates genes that sense DNA damage, such as the p53 tumor suppressor gene, which would otherwise trigger apoptosis in response to the physiologic DNA rearrangements of the germinal center, and also inhibits terminal B cell differentiation. This is described in more detail separately. (See "Pathobiology of diffuse large B cell lymphoma and primary mediastinal large B cell lymphoma", section on 'Aberrant BCL6 expression'.)

Other mechanisms — Because BCL-2 overexpression is not sufficient for FL development, other genetic lesions or host factors must be required. The need for additional factors is supported by the observation that BCL-2 transgenic mice develop polyclonal hyperplasia of mature, long lived non-dividing B cells rather than a clonal proliferation resembling FL [5,6]. With time, and analogous to the human disease, a fraction of BCL-2 transgenic mice develops aggressive, clonal large cell lymphomas, which have acquired additional genetic lesions [29]. Clues to mechanisms beyond BCL-2 overexpression that are involved in the development of FL have emerged from deep sequencing of FL genomes and studies of the FL microenvironment. (See 'Tumor microenvironment' below.)

Other genetic lesions — Approximately 90 percent of FL demonstrate genetic alterations in addition to the (14;18) translocation [1]. The most common are loss of 1p, 6q, 10q, and 17p and gains of chromosomes 1, 6p, 7, 8, 12q, X, and 18q/dup. As examples:

●Chromosome 6 – Deletions in the long arm of chromosome 6 occur in approximately 20 percent of cases and have been associated with worse prognosis in multiple studies [30-32].

●Chromosome 1 – 1p36, the site of the TNFRSF14 gene [33], is affected by deletions, uniparental disomy, and point mutations in a subset of cases, and was previously associated with poor prognosis.

●Chromosome 3 – About 5 to 15 percent of FLs have 3q27 abnormalities involving the BCL-6 gene, which is commonly involved by genetic aberrations that contribute to the pathogenesis of diffuse large B cell lymphoma [27,28] (see 'BCL-2 negative cases' above). By contrast, other proto-oncogenes involved in lymphomagenesis (eg, MYC, cyclin D1) are not involved in FL.

Efforts focused on genome sequencing have identified a role for mutation of histone-modifying genes and chromatin-modifying genes in FL. Multiple studies have reported mutations in the following genes (listed in order of decreasing frequency): KMT2D, CREBBP, EZH2, EP300, HIST1H1E, KMT2C, ARID1A, and SMARCA4 [34-42]. The mechanism by which these genes promote FL remains to be elucidated, but work looking at the hierarchy of these genetic aberrations has suggested that BCL-2 rearrangements, CREBBP mutations, and mutations increasing the function of EZH2 are early "founding" events, and that KMT2D and TNFRSF14 mutations occur later in the course, possibly as part of disease progression and/or development of drug resistance [34,37,43]. Disruption of KMT2D perturbs germinal center B cell development and accelerates the onset of germinal center B cell derived lymphomas in BCL-2 transgenic mice, confirming that loss of KMT2D activity is lymphomagenic [44]. In addition, several studies have identified recurrent mutations in histone genes [38,39,45,46], reinforcing the notion that abnormalities of chromatin have an important, albeit incompletely understood, role in the pathobiology of FL.

Other sequencing studies have identified recurrent mutations involving signaling pathways with important roles in B cell biology. One group of recurrent mutations is hypothesized to upregulate mTOR (mammalian target of rapamycin) signaling in FL cells. These mutations, which are found in approximately 15 to 20 percent of cases, may involve genes encoding components of vacuolar H⁺-ATPase, which is known to be necessary for amino acid-induced mTOR activation, or RRAGC, a guanine nucleotide binding protein that also regulates mTOR activation [47]. Somatic mutations also have been identified in a subset of FL in genes that encode components of the B cell receptor (BCR) signaling pathway (eg, BTK, CD79B) [45], which controls many facets of B cell biology, including growth, differentiation, survival, and adhesion or cellular migration.

Acquired genetic lesions generally increase in number with increasing FL histologic grade, being most numerous with grade 3b histology. Over time, a significant fraction of FL evolves into an aggressive lymphoma, most commonly with a diffuse large cell histology [48]. Morphologic transformation is often associated with disruptive mutations in the tumor suppressor gene TP53 on chromosome 17p or the tumor suppressor locus CDKN2A/2B on chromosome 9, which encodes a negative regulator of cell cycle progression and a stabilizer of p53 [38,49-51]. Other common events associated with transformation include amplifications, translocations, and mutations of MYC, an important regulator of cell growth, and loss-of-function mutations in B2M that abolish the expression of major histocompatibility class I antigens, indicating the immune evasion contributes to histologic translocation in some cases [38]. (See "Histologic transformation of follicular lymphoma", section on 'Cytogenetics'.)

Tumor microenvironment — The tumor microenvironment, specifically the immunologic microenvironment comprised of T cells and dendritic cells, may influence the development and progression of FL [52-56]. It has been suggested that FL is an immunologically functional disease in which an interaction between the tumor cells and the microenvironment determines overall clinical behavior.

The role of the microenvironment is supported by gene expression profiling and immunophenotypic studies that have correlated patient survival with types of immune cells that are intermixed with the lymphoma, rather than with the gene expression program of the lymphoma itself [57-60]. Communication between the tumor cells and the microenvironment appears to be bidirectional and likely involves chemokines, chemokine receptors and adhesion molecules, the balance of which determines whether tumor cell growth is stimulated or inhibited [61,62]. Signals from the tumor cells help to determine the composition of the cells in the surrounding area, which in turn support the growth and survival of the malignant cells [63,64].

Comparisons of gene expression between purified FL cells and normal germinal center B cells have found upregulation of cell cycle mediators, B cell transcription factors, certain cytokines, and genes involved in cell-cell interactions [65]. These differences are consistent with a role for the microenvironment in tumor pathogenesis that requires further study.

In some instances, mutations involving FL cells may directly influence the nature of their interaction with the microenvironment. FL cells retain expression of surface immunoglobulin (Ig, also known as B cell receptor or BCR), and evidence suggests that signals mediated by the BCR are critical for FL growth and survival. Of interest, the Ig genes in FL in a high fraction of cases acquire mutations that introduce consensus sequences for N-glycosylation, resulting in an unusual post-translational modification in which the added glycan is highly mannosylated [66]. This modified BCR binds a lectin called dendritic cell-specific ICAM3-grabbing nonintegrin (DC-SIGN), which is expressed on infiltrating macrophages and initiates BCR signaling [67,68]. In a second example, a subset of FL has gain-of-function mutations in the CTSS gene, which encodes the protease cathepsin S [69]. Increases in cathepsin S function appear to enhance HLA class II presentation of antigen, leading to increases in the activation of CD4-positive T cells, recruitment of CD4-positive T cells, release of pro-inflammatory cytokines, and growth of FL cells. (See "Clinical manifestations, pathologic features, diagnosis, and prognosis of follicular lymphoma", section on 'Microenvironment'.)

HETEROGENEITY OF FL

Spacial and temporal heterogeneity — Within an individual, there are differences in the FL tumor cells and tumor environment found at different disease sites at the same time (spacial heterogeneity) and at the same disease site across time (temporal heterogeneity).

●Spacial heterogeneity – Biopsies of multiple tumor sites reveal genomic differences with varying patterns of somatic mutations and gene expression across sites within an individual [70,71]. While our understanding about the clinical impact of these differences is evolving, some may help explain discordant responses to therapy at different disease sites.

●Temporal heterogeneity – FL tumor cells within individual patients are genetically heterogeneous, with several subpopulations of malignant cells evolving over time [41]. The relative contribution of each subpopulation to the total tumor volume is likely affected by interactions with other subpopulations, the tumor microenvironment, and therapy. This heterogeneity represents genetic variation upon which selection can act and contributes to the development of resistant disease and progression (transformation) to more aggressive disease. (See 'Transformation' below.)

Diversity of cases — The diagnostic category of FL is heterogeneous in terms of morphology, genetics, and biologic behavior. It is generally unclear how this variation relates to clinical outcome and response to therapy, which constitute areas of ongoing investigation.

As examples:

●Tumors with EZH2 mutations show increased H3K27 trimethylation [72] and have a better prognosis than EZH2 wild type FLs [73,74]. The increase in H3K27 trimethylation in EZH2 mutated FL is a direct result of gain-of-function mutations in EZH2 [37], an insight that has led to use of the EZH2 inhibitor tazemetostat for treatment of relapsed/refractory FL with EZH2 mutations [75].

●FL that develops in the setting of hepatitis B virus infection is less likely to be associated with the t(14;18) and appears to have a more aggressive clinical course [76]. This subgroup of FL have mutational and gene expression signatures and microenvironmental features that are distinct from FL developing in the absence of hepatitis B virus infection.

●Full understanding of the FL microenvironment requires characterization of not only FL cells and reactive immune cells, but also non-hematopoietic cells such as follicular dendritic cells, other fibroblast-derived stromal cells, and blood vessel- and lymphatic vessel-associated cells. Studies using single cell RNA-sequencing have begun to reveal changes in non-hematopoietic cells in the FL microenvironment, a first step in detailing crosstalk between FL cells and non-hematopoietic cells that promotes FL growth and survival [77].

TRANSFORMATION — Histologic transformation (HT) refers to the evolution of a clinically indolent non-Hodgkin lymphoma (NHL; eg, FL) to a clinically aggressive NHL (eg, diffuse large B cell lymphoma [DLBCL]) defined as those lymphomas in which survival of the untreated patient is measured in months. This is discussed in more detail separately. (See "Histologic transformation of follicular lymphoma".)

SUMMARY

●Cell of origin – Follicular lymphoma (FL) is the second most common histologic subtype of non-Hodgkin lymphoma. The molecular pathogenesis of FL is a complex, multistep process leading to the outgrowth of a malignant clone of germinal B cell origin (figure 1). While some steps in this pathway have been elucidated, many remain unknown. (See 'Cell of origin' above.)

●Aberrant BCL-2 expression – The development of the majority of FL tumors in adults is associated with the overexpression of B cell leukemia/lymphoma 2 (BCL-2) located on chromosome band 18q21. BCL-2 is an oncogene that blocks programmed cell death (apoptosis). As such, overexpression results in prolonged cell survival. (See 'Aberrant BCL-2 expression' above.)

In the vast majority of cases, FL is associated with a translocation between the long arm of chromosome 18, the site of the BCL-2 oncogene, and one of the three immunoglobulin (Ig) genes. The most common translocation involves the Ig heavy chain gene resulting in the t(14;18)(q32;q21), found in approximately 85 percent of FL. (See 'Translocations involving BCL-2' above.)

●BCL-2 negative cases – Less than 10 percent of FL tumors lack BCL-2 translocations and fail to express BCL-2 proteins. BCL-2 negativity is most commonly seen in grade 3b FL, in which the tumor is comprised of solid sheets of centroblasts that nonetheless grow in a follicular pattern. (See 'BCL-2 negative cases' above.)

●Other mechanisms – BCL-2 overexpression in itself is not sufficient for FL development and other genetic lesions or host factors are required. Many of the complementary mutations involve genes that regulate the epigenome (eg, histone modifications, chromatin structure). Other important factors involve the local host response and tumor microenvironment, and modifications in the B cell receptor that may augment B cell receptor signaling. (See 'Other mechanisms' above.)