AUTHORS
Sophia Chaudhry, PhD, Samantha J Lamantia, MS2, Wenjie Sun, PhD3, Derek H. Janssens, PhD4, Tim Triche Jr, PhD5
AFFILIATIONS
1 Van Andel Institute, 333 Bostwick Ave NE, Michigan, United States
2 Van Andel Institute, 333 Bostwick Ave NE, Michigan, United States
3 Van Andel Institute, 333 Bostwick Ave NE, Michigan, United States
4 Van Andel Institute, 333 Bostwick Ave NE, Michigan, United States
5 Van Andel Institute, 333 Bostwick Ave NE, Michigan, United States
ABSTRACT
DNA-binding transcription factors are overexpressed in numerous hematological malignancies, but how they contribute to oncogenic transformation and whether these proteins are acting in a conserved or context-specific manner that may contribute to tumor heterogeneity remains unclear. The t(14;16)(q32;q23) translocation in multiple myeloma (MM) juxtaposes the IGH promoter to the c-MAF gene, yielding constitutive overexpression. Previous c-MAF profiling studies were applied to models with additional genomic alterations such as MYC rearrangements. It is unclear how c-MAF directs oncogenic gene regulatory changes and which of these alterations may depend on interactions with the commonly amplified cofactor MYC. Therefore, we compared c-MAF occupancy across cell line models with t(14;16)(q32;q23) translocations alone, or in combination with MYC amplification. We identified c-MAF targets expressed in normal plasma cells, as well as context-specific programs of c-MAF activation that suggest a potential role in promoting MM remodeling of the bone-marrow microenvironment.
CUT&Tag was used to profile c-MAF, H3K4me2, H3K4me3, and H3K27me3 in ANBL6 cells, a t(14;16) model with no secondary MYC rearrangements. For comparison, c-MAF ChIP-seq profiles from MM.1S and JJN3 (14;16/c-MAF + 8;14/MYC) were obtained from public datasets. ANBL6 c-MAF peaks that were found in all three models were classified as core-conserved. We also identified c-MAF peak sets shared by ANBL6 and MM.1S, ANBL6 and JJN3, or specific to ANBL6.
Among the 11,481 ANBL6 c-MAF peaks, 3,059 (26.6%) were core-conserved, 1,236 (10.8%) were MM.1S-shared, 1,365 (11.9%) were JJN3-shared, and 5,821 (50.7%) were ANBL6-specific, demonstrating that approximately half of ANBL6 c-MAF occupancy was model-restricted. These core-conserved sites overlapped with active chromatin features marked by H3K4me2 more frequently than ANBL6-specific sites (54.2% vs. 29.1%, p<0.001) while core-conserved sites overlapped with active promoters marked by H3K4me3 less frequently than ANBL6-specific sites (34.8% vs. 44.5%, p<0.001). In addition, core-conserved sites overlapped with repressed chromatin features marked by H3K27me3 less frequently than ANBL6-specific sites (3.8% versus 17.7%, p<0.001). Motif enrichment distinguished the four peak classes. Core-conserved sites contained c-MAF/bZIP recognition sequences and RUNX1, IRF4, and BACH2 family motifs, whereas ANBL6-specific sites were enriched for candidate motifs IRF/ISRE, STAT, SMAD, and IKZF. Together, this indicates the context-specific cMAF binding sites are enriched over promoters and regions of repressive chromatin and may be directed through interactions with immune regulatory proteins.
We next compared the ANBL6 c-MAF peaks to the chromatin landscape of normal plasma cell states. Among the 305 high-confidence genes associated with c-MAF and H3K4me3 co-occupancy, 223 (73.1%) are already active in healthy plasma cells, consistent with c-MAF reinforcing established plasma-cell transcriptional programming. However, we identified a smaller candidate program of 23 genes (7.5%) repressed in healthy plasma cells, indicating c-MAF is involved in promoting ectopic activation of these genes in MM cells. We confirmed the expression of 14/23 of these genes by RNA-seq expression analysis. This group of c-MAF-activated targets include the perivascular-associated genes RGS5, FGF2, and ALDH1L1, and DKK1, and the inflammatory regulators CLEC7A and STAT4.
Here, we define a conserved set of c-MAF-activated sites in t(14;16) MM and reveal context-specific roles of c-MAF that may contribute to oncogenic transformation and tumor heterogeneity. The core-conserved regions indicate that c-MAF reinforces established plasma cell programming, and that the context-specific sites that c-MAF binds, independently of MYC amplification, are associated with activation of a smaller niche-associated, perivascular, and inflammatory program. This candidate program indicates that c-MAF potentially reprograms the bone-marrow microenvironment to be more advantageous for plasma cell survival, providing a new avenue for targeting c-MAF-driven MM in the future.
DISCLOSURES
Sophia Chaudhry, PhD: No relevant conflicts of interest to declare.