Comparative analysis of macrophage feeder systems reveals distinct behaviors and key transcriptional shifts in chronic lymphocytic leukemia cells via coculture

The development and progression of chronic lymphocytic leukemia (CLL) are driven not only by the intrinsic properties of leukemia cells but also by their complex interactions with the tumor microenvironment.1 The myeloid compartment, particularly macrophages, plays a crucial role in driving CLL progression and therapy resistance.2, 3 Blood monocytes differentiate in vitro under the influence of CLL cells into nurse-like cells (NLCs), which protect leukemic cells from spontaneous apoptosis4 and promote multidrug resistance.5 Macrophage depletion in vivo using CSF1R blockade or liposomal clodronate significantly reduced leukemic burden, demonstrating their important role in CLL pathogenesis.6, 7 However, our understanding of the precise mechanisms by which macrophages promote CLL survival remains incomplete. To dissect the molecular dialog between CLL cells and macrophages, both in vivo models and controllable in vitro systems are essential. Although some macrophage–CLL coculture systems exist,4, 8-10 systematic, simultaneous analyses of these systems are lacking. Therefore, we evaluated various macrophage coculture systems for their CLL-feeding potential, phagocytosis capacity, induction of treatment resistance, and their impact on CLL transcriptional profiles (Figure 1A). Several human and mice macrophage systems were used. Human systems included THP-1 macrophages10 differentiated with phorbol-12-myristate-13-acetate, healthy donor monocyte-derived macrophages (HD-MDM) differentiated from peripheral blood mononuclear cells (PBMCs), and NLCs generated from CLL PBMCs. NLC purity was confirmed by flow cytometry and microscopy (Supporting Information: Figure S1). Murine systems included primary bone marrow-derived macrophages10 (BMDMs), and J774A.18 and MacCsf1r+/+ macrophage11 cell lines. Whereas primary and THP-1 macrophages do not proliferate after differentiation, J774A.1 cells show robust proliferation and phagocytosis. Thus, J774A.1 macrophages were γ-irradiated to halt proliferation. All macrophage systems were cultured simultaneously with eight treatment-naïve CLL samples (Supporting Information: Table S1). CLL viability was measured on Days 0, 1, 3, 5, and 7 by flow cytometry. All macrophage systems significantly supported CLL viability throughout the 7-day period (Figure 1B) despite interpatient variability (Supporting Information: Figure S2A). Due to the lower NLC count, CLL viability was the lowest in NLC, but this difference narrowed considerably (Supporting Information: Figure S2B) when all macrophage systems were seeded at the same density as the average NLC count (Supporting Information: Table S3). Moreover, fresh and thawed CLL cells showed no significant difference in the viability in cocultures (Supporting Information: Figure S3A), and the coculture systems did not induce CLL proliferation (Supporting Information: Figure S3B). Interestingly, autologous versus allogenic CLL–NLC pairs showed no difference in viability support (Supporting Information: Figure S4), and M2 polarization of macrophages did not enhance CLL viability beyond M0 levels (Supporting Information: Figure S5). To assess spontaneous phagocytic capacity, all systems were cocultured with CLL cells for 18 h, and the remaining CLL cells were counted. While HD-MDMs, NLCs, and THP-1 macrophages showed modest phagocytosis, murine macrophages showed high rates, with non-irradiated J774A.1 cells demonstrating the highest rate, followed by MacCsf1r+/+ and BMDM (Figure 1C), implying that CLL cell clearance by murine macrophages should be particularly considered. To assess antibody-dependent cellular phagocytosis (ADCP), another key function of macrophages, CLL–macrophage cocultures were exposed for 18 h to the monoclonal anti-CD20 antibody obinutuzumab. We observed no significant difference in ADCP between the systems, except for irradiated J774.A1 (Figure 1D). When combined with the higher spontaneous phagocytosis of murine macrophages, these results suggest that human macrophages are more effective in ADCP than murine ones. Altogether, these assays highlighted fundamental differences between human and murine macrophages and the importance of species-specific immune cell interactions. Emerging evidence suggests that macrophages significantly influence treatment outcome and therapy response of CLL patients.3 To evaluate the suitability of the coculture systems in drug-testing experiments, we treated human macrophage–CLL cocultures with the BCL2 inhibitor venetoclax and the BTK inhibitor ibrutinib. High-dose ibrutinib (10 µM) significantly reduced the CLL cell survival cocultured with HD-MDMs or NLCs, whereas THP-1 coculture appeared to confer resistance to ibrutinib-induced CLL apoptosis, showing no remarkable decrease in viability (Figure 1E). Both low- and high-dose venetoclax were potent in killing most CLL cells in mono- and cocultures. However, THP-1 coculture again provided partial protection, resulting in higher CLL cell viability than other systems (Figure 1E). Aside from reduced HD-MDM viability at 10 µM ibrutinib, macrophage layers were not markedly affected by drug exposure (Supporting Information: Figures S6). Our finding demonstrates a more significant protection of THP-1 feeder for venetoclax-treated CLL cells than previously reported,9 suggesting that larger cohorts are needed for a definitive conclusion. Current studies have shown promising efficacy for the bispecific antibody epcoritamab for refractory patients.12, 13 We tested epcoritamab in a mixed culture of CLL-PBMCs with the human macrophage feeders. While HD-MDMs and autologous NLCs showed reduced CLL cell viability, indicating an epcoritamab-dependent T cell-mediated killing, THP-1 cells abrogated this effect, possibly due to the tumor origin of this cell line (Figure 1F). To investigate the molecular mechanisms by which macrophages affect CLL cells, sorted CLL cells (Supporting Information: Figure S7) of three patients after 5 days of coculture with HD-MDMs, THP-1 macrophages, NLCs, and BMDMs were collected and subjected to bulk mRNA sequencing. CLL cells cultured with BMDMs showed minimal transcriptional changes (Supporting Information: Figure S8A), whereas coculture with human macrophages led to significant and distinct gene expression shifts: 192 with NLCs (Supporting Information: Figure S8B), 660 with THP-1 (Supporting Information: Figure S8C), and 263 with HD-MDMs (Supporting Information: Figure S8D), highlighting feeder-specific effects on CLL cells. Gene set enrichment analysis using the hallmark collection from the Molecular Signatures Database identified eight hallmarks enriched in all three human systems (Figure 2A). Notably, “inflammatory response,” “IL2 STAT5 signaling,” “IL6 JAK STAT3 signaling,” and “KRAS signaling” were consistently upregulated in CLL cells upon coculture. Focusing on the crucial signaling pathways in B cell biology, we used a curated set of gene signatures related to signaling pathways, transcription factor regulation, and key cellular processes such as proliferation and metabolism known to influence CLL pathobiology (Supplemental Gene Set). Again, this analysis revealed upregulation of “STAT3,” “IL6,” “RAS,” and “Proliferation” across all systems (Figure 2B). JAK signaling was also elevated after HD-MDM- and NLC coculture. Additionally, pathway enrichment analyses using the Gene Ontology and Reactome databases also highlight enhanced ERK/MAPK signaling in CLL cells after cocultures (Supporting Information: Figure S9). Altogether, these analyses highlighted the IL6/JAK/STAT3 and RAS/MAPK pathways as potential key mechanisms through which macrophages may drive CLL cell survival. In CLL, constitutive activation of NFκB signaling in CLL cells induces IL6 production, which in turn activates the JAK/STAT pathways, creating a feed-forward loop that promotes leukemogenesis.14 We could confirm the activation of the JAK/STAT3 pathway by immunoblotting of independent CLL samples after cocultures (Figure 2C and Supporting Information: Figure S10), showing increased phospho-STAT3 and phospho-JAK1 across patients in all cocultures. Phospho-JAK1 and total JAK2 levels varied strongly across patients. Elevated RAS signaling was identified in all cocultured CLL cells, revealing another pathway involved in macrophage-mediated support. Heightened RAS/MAPK activation, often driven by RAS mutations in CLL, correlates with adverse clinical features and shorter treatment-free survival.15, 16 Immunoblotting also confirmed strongly enhanced phosphorylation of the RAS mediators MEK, ERK, and p38 MAPK across patients in THP-1 and HD-MDM cocultures, despite the absence of enhanced BCR activation based on phospho-SYK and phospho-BTK levels (Supporting Information: Figure S10), suggesting alternative upstream activation of ERK signaling such as RAS. Future studies should test whether JAK/STAT or MAPK blockade may disrupt CLL survival support. While earlier findings have noted BCR activation upon NLC contact,17 BCR signaling was downregulated in CLL cells cocultured with HD-MDMs and NLCs in our GSEA and in THP-1 coculture in immunoblot (Figure 2B and Supporting Information: Figure S10). This is further confirmed by our extended analysis across multiple time points and culture conditions (Supporting Information: Figure S11A), confirming enhanced BCR activation of CLL cells during the NLC generation phase (Supporting Information: Figure S11B), but unchanged or even reduced phospho-SYK and phospho-BTK during cocultures of allogeneic CLL cells with the fully differentiated macrophages (Supporting Information: Figure S11C). This finding illustrates the context-dependent nature of BCR regulation in CLL, and indicates that survival support provided by fully differentiated macrophages can occur independently of BCR activation. The CLL lymph node environment was shown to enhance the “Proliferation” signature, associated with increased BCR, NFκB, and NOTCH signaling in CLL cells.18-20 Although we detected elevated NFκB and NOTCH signaling in CLL cells cocultured with HD-MDMs and THP-1 macrophages (Figure 2A,B), the lack of BCR activation and leukemic proliferation in our systems indicates that macrophage cocultures are insufficient to induce the prominent lymph node signatures, underscoring the collaborative role of different microenvironmental components. Collectively, all tested macrophage feeders effectively maintained CLL cell viability in vitro, despite notable differences in phagocytosis, ADCP capacity, or drug responses. Our transcriptomic and protein-level analyses revealed that human macrophages prominently upregulate JAK/STAT and MAPK pathways in CLL cells. Given the complexity of the CLL microenvironment and the pivotal role of macrophages in disease pathogenesis, our study highlights the importance of selecting an appropriate coculture system to obtain reliable mechanistic insights. THP-1 macrophages induced the strongest transcriptional changes in the CLL cells and conferred resistance to inhibitors, warranting caution in drug-testing applications. Notably, HD-MDMs closely mirrored NLC behaviors across our assays, supporting their use as a more accessible alternative model. Cross-species systems should be interpreted with caution due to elevated spontaneous phagocytosis. Finally, future studies examining macrophage polarization and the feeder system-specific differences in gene expression and cytokine production can further refine model selection and mechanistic understanding. We thank Daniel Bachurski, Michael Michalik, and the CLL Biobank Cologne (University of Cologne) for providing materials and patient samples and E. Richard Stanley (Albert Einstein College of Medicine) for providing the MacCsf1r+/+ cell line. Figure 1A and Supporting Information: Figure S11A were created with BioRender.com. Open Access funding enabled and organized by Projekt DEAL. Viktoria Kohlhas: Conceptualization; investigation; funding acquisition; writing—original draft; methodology; writing—review and editing; project administration; visualization; data curation. Hendrik Jestrabek: Investigation; methodology; writing—original draft; writing—review and editing; visualization; data curation. Rocio Rebollido-Rios: Investigation; writing—review and editing; formal analysis; data curation; visualization. Thanh Tung Truong: Investigation; writing—review and editing. Rebekka Zölzer: Investigation; writing—review and editing. Luca D. Schreurs: Visualization; investigation; writing—review and editing. Duc Pham: Investigation; writing—review and editing. Alexander F. vom Stein: Visualization; writing—review and editing; investigation. Michael Hallek: Resources; supervision; writing—review and editing; funding acquisition. Phuong-Hien Nguyen: Conceptualization; funding acquisition; writing—original draft; writing—review and editing; methodology; project administration; supervision; resources. The authors declare no conflict of interest. All ethics approvals and patient consents are provided in the manuscript. This study is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) grant SFB1530–455 784 452 to P.-H.N and M.H. (sub-project B01) and V.K. (postdoc advancement grant), and by the research network CANcer TARgeting (CANTAR) of the Ministry of Culture and Science of the State of North Rhine-Westphalia (MKW NRW) to P.-H.N. and M.H. The data that support the findings of this study are openly available in the BioStudies database (http://www.ebi.ac.uk/biostudies) at https://www.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-15 353. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.