Skip to Main Content


Harnessing negative B-cell selection to overcome drug-resistance in B-cell malignancies

Figure 1: Leveraging mechanisms of negative B-cell selection to overcome drug-resistance.

Thresholds of B-cell selection based on signaling strength of the BCR (left) or an oncogenic mimic (BCR-ABL1; right) are depicted. BCR-signals are amplified by kinases (e.g. SYK, PI3K) and attenuated by inhibitory phosphatases (e.g. SHIP1, PTEN) to achieve intermediate signaling strength (middle).
When B-cells lack BCR-signaling (Anergic; bottom left) or in the presence of kinase-inhibitor treatment in B-cell malignancies (bottom right), signaling strength falls below minimum levels for proliferation and survival.
Conversely, upon BCR-hyperactivation in autoreactive B-cells (top left) or kinase-hyperactivation in B-cell malignancies (e.g. by kinase-superagonists or small-molecule phosphatase-inhibitors, top right), signaling strength exceeds maximum thresholds and triggers negative selection and cell death.

To prevent the production of harmful autoantibodies and autoimmune disease, autoreactive B-cells are eliminated by a process we termed negative selection. Contrary to established dogma, these mechanisms are not only active in preventing autoimmune disease but as demonstrated by our work (Trageser et al., 2009; Swaminathan et al., 2013; Chen et al., 2015; Shojaee et al., 2016), they also represent an entirely novel class of therapeutic targets in B-cell tumors, including B-cell acute lymphoblastic leukemia (B-ALL), chronic lymphocytic leukemia (CLL) and mantle cell lymphoma (MCL).

Despite oncogenic transformation, these B-cell malignancies remain fully sensitive to mechanisms of negative selection. Targeted hyperactivation of B-cell receptor (BCR)-downstream kinases mimics excessive signaling-strength from an autoreactive BCR and triggers negative selection. This can be achieved by pharmacological kinase hyperactivation using superagonists of SYK and other BCR-proximal kinases (Figure 1). Targeted cancer-therapy is traditionally focused on kinase-inhibitors to suppress oncogenic signaling. Their concept of targeted kinase-hyperactivation effectively represents the opposite (Müschen, 2018; Müschen, 2019)

Feedback control of oncogenic signaling in B-cell malignancies as a therapeutic target

Unlike other cell-types, B-cells rely on a ‘Goldilocks’ principle that selects B-cell clones for survival and proliferation based on intermediate B-cell receptor (BCR) signaling strength. In analogy to normal B-cells, we discovered that in many B-cell malignancies, the transforming oncogene mimics a constitutively active (pre-)BCR (Feldhahn et al., 2005).

Based on these observations, we demonstrated that B-cell acute lymphoblastic leukemia (B-ALL), chronic lymphocytic leukemia (CLL) and mantle cell lymphoma (MCL) cells critically depend on robust feedback mechanisms to balance fluctuations of oncogenic BCR-signaling strength (Buchner et al., 2015: Shojaee et al., 2015; Chan et al., 2020; Lee et al., 2020). The efforts for preclinical development of this concept focused on small-molecule inhibition of DUSP6, a central feedback regulator of ERK-signaling.

Metabolic gatekeeper function of B-cell transcription factors

While B-cell transcription factors PAX5 and IKZF1 are known for their role in B-cell differentiation, we recently discovered a novel B-lymphoid transcriptional program to restrict energy abundance (Chan et al., 2017), a mechanism named ‘metabolic gatekeeper’ (Müschen, 2019). Oncogenic signaling and increased proliferation results in higher energy demands of transformed B-cells than their normal counterparts.

Hence, B-cell transcription factors function as metabolic gatekeepers (Chan et al., 2017; Schjerven et al., 2017; Xiao et al., 2018; Sadras et al., 2021; Pan et al., 2021) by limiting energy-supply and safeguarding against malignant transformation by setting low thresholds for the elimination of premalignant clones based on energy-stress.

BCL6 enables stemness and a novel form of drug-resistance in B-cell leukemia

While B-cell lymphoma 6 (BCL6) is widely known as oncogene in mature B-cell lymphomas, we discovered that B-cell acute lymphoblastic leukemia (B-ALL) cells respond to treatment with tyrosine kinase inhibitors (TKI) by over 100-fold increases of BCL6-expression, which drives a previously unrecognized form of drug-resistance (Duy et al., 2011; Nahar et al., 2011; Hurtz et al., 2011).

For mechanistic studies, we developed genetic mouse models for conditional ablation of Bcl6 and a Bcl6-reporter allele (Bcl6tm1Mamu) and found that BCL6 mediates drug-resistance by induction of a quiescent stem cell-like state in B-ALL and other leukemias. Our work demonstrated that BCL6 peptide- and small-molecule antagonists restored responses to standard chemotherapy and TKI in refractory leukemias (Hurtz et al., 2019).

Immunoglobulin diversifiers in clonal evolution of B-cell precursors towards B-ALL and lymphoma

Unlike other cell-types, B-cells undergo multiple rounds of gene-recombination (mediated by the RAG1 and RAG2 enzymes) and hypermutation (mediated by AID) to evolve high-affinity antibodies. High frequencies of DNA-strand breaks engender an approximately 300-fold increased risk of malignant transformation. While AID and RAG1/2 are strictly segregated during normal B-cell development, we found that their aberrant co-expression plays a central role in clonal evolution towards leukemia (Feldhahn et al., 2007; Tsai et al., 2008; Klemm et al., 2009; Kharabi Masouleh et al., 2014; Swaminathan et al., 2015; Huang et al., 2019 ).

In addition, we demonstrated that AID enables clonal evolution towards B-cell acute lymphoblastic leukemia (B-ALL) and mantle cell lymphoma (MCL) relapse and drug-resistance (Klemm et al., 2009).

Approaches to Target Discovery

To generate hypotheses for target discovery, we build on clinical outcome predictors: in collaboration with multiple study groups across the US, we have developed and validated phenotypic biomarkers of favorable and poor clinical outcomes in B-cell acute lymphoblastic leukemia (B-ALL), chronic lymphocytic leukemia (CLL) and mantle cell lymphoma (MCL) and integrated these markers into models of oncogenic signaling pathways (Chan et al., 2017).

As part of the NCI CTEP ‘Human Hematopoiesis and Leukemia PDX’ program, our lab developed patient-derived xenografts resources to model B-cell malignancies based on patient-derived cells and cord blood-based humanized mouse models to study mechanisms of human B-lymphopoiesis in vivo. To identify potential B-cell-specific vulnerabilities, we have developed a bioinformatic target-discovery resource (Müschen, 2018; Müschen, 2019) based on mutation and gene expression data from over 160 clinical trials, covering 39 cancer types and more than 25,000 patients. To exemplify the usefulness of this platform, an integrated analysis of these data sets led to the discovery that common genetic lesions promoting PI3K-activation in cancer were strongly counterselected in 2,375 B-ALL, CLL and MCL cases studied.

Functional follow-up studies revealed that these B-cell malignancies are uniquely sensitive to PI3K-hyperactivation, an unexpected vulnerability that we are currently pursuing for therapeutic targeting. Based on recent discoveries, the Müschen lab will focus on a new ‘B-cell selection paradigm’ for the treatment of B-cell malignancies to target thresholds of oncogenic signaling-strength instead of traditional approaches that are exclusively focused on suppression of oncogene activity.

Design of kinase-superagonists to induce negative B-cell selection

We discovered multiple drug-targets for this approach, including inhibitory phosphatases that balance oncogenic signaling strength in B-cell tumors. Small molecule-inhibitors of the SHIP1- and PTEN phosphatases have shown the greatest promise so far (Chen et al., 2015; Shojaee et al., 2016). In addition, we identified SYK-hyperactivation at the apex of the BCR-signaling cascade as a central driver of negative selection.

Hence, our lab is pursuing the design and development of SYK-superagonists that lock SYK in an open conformation to break constitutive autoinhibition of this kinase.

Leveraging B-cell selection to short-circuit clonal evolution and relapse

Figure 2: Sequential kinase-inhibitor and superagonist-treatment to subvert clonal evolution and relapse.

Combinations of agents applying selective pressure in the same direction (e.g. kinase-inhibitors) select for clones based on kinase-inhibitor resistance (left panel).
By contrast, kinase-inhibitors and superagonists apply selective pressures on malignant B-cell clones in opposite directions. A sequential treatment regimen that alternates between kinase-inhibitors and super-agonists will compromise Darwinian selection for drug-resistant mutants (right panel).
In this case, selection for kinase-inhibitor resistance in the first round of treatment (kinase-inhibitor) will not increase adaptive fitness in the second round of treatment when selective pressure is turned in opposite direction (kinase-superagonists).

Despite substantially improved outcomes, drug-resistance and relapse of refractory disease remain central problems in the treatment of patients with B-cell malignancies. Current treatment regimens use agents that apply selective pressure in only one direction (kinase-inhibition). Contrary to this tenet, we propose the new concept based on sequential treatment regimens that alternate between kinase-inhibitors and superagonists.

By sequentially applying selective pressures in opposite directions, this approach will subvert Darwinian selection for drug-resistant mutants (Figure 2).

While one-directional treatment-regimens are effective in solid tumors and myeloid disorders, this bi-directional approach harnesses the Goldilocks-nature of B-cell-selection to subvert mechanisms of clonal evolution (Swaminathan et al., 2015).

Why are B-cells uniquely sensitive to kinase-hyperactivation?

Macrophages are not at risk to express autoantibodies and, hence, not subject to negative selection. To test B-cell-specificity, we overexpressed the transcription factor CEBPα to reprogram B-cells into macrophages, which abolished negative selection and sensitivity to kinase-hyperactivation (Chen et al., 2015; Shojaee et al., 2016).

Hence, we will explore why B-cells, but not other cell-types are sensitive to kinase hyperactivation. This work will assume the new concept that B-cell selection is driven by lineage-specific regulation of energy abundance, a mechanism termed ‘metabolic gatekeeper’.

Metabolic gatekeepers as a single mechanism to clear autoreactive and (pre-)malignant B-cells

Figure 3: Metabolic gatekeepers set thresholds for negative selection of autoreactive and premalignant clones.

In the presence (a) but not in the absence (b) of functional metabolic gatekeepers, glucose and glutamine uptake is limited to levels that are insufficient to fuel energy demands of an autoreactive BCR or an oncogenic kinase. Gatekeeper functions arefrequently compromised by genetic deletion of PAX5 or IKZF1 (b). In the absence of metabolic gatekeepers, ATP levels are no longer restricted and energy abundance is sufficient to fuel increased energy demands downstream of an autoreactive BCR or a transformaning oncogene.

In the presence of intact metabolic gatekeepers, prolonged kinase-hyperactivation, downstream of an autoreactive BCR or an oncogene, will exhaust ATP-reserves, resulting in energy-stress and negative selection (Figure 3).

Since no other cells produce autoantibodies, we hypothesize that mechanisms to remove autoreactive B-cells are not only driven by B-cell transcription factors (PAX5, IKZF1) but also determined by features that distinguish B-cells from other cells, namely smaller size, fewer mitochondria, and a faster cell division-rate than any other cell-type.

Recent findings by the Müschen laboratory suggested that B-cells are, by default, in a state of chronic energy-depletion as a result of unique morphological, transcriptional and biochemical characteristics (Chan et al., 2017; Schjerven et al., 2017; Xiao et al., 2018; Sadras et al., 2021; Pan et al., 2021).