(690c) Targeting the Pro-Lymphoma Stromal Microenvironment

The lymphoma microenvironment (LME) is inherently dynamic and complex, consisting of cellular as well as non-cellular compartments. These compartments (i.e. malignant, stromal, immune) facilitate a heterogeneous physico-chemical dialogue that often drives the progression of lymphoma, and subsequent resistance to treatments. Data from animal models suggest that malignant cells are able to hijack and corrupt the normal physiology of stromal cells, thereby rendering them pro-tumorigenic. However, the key effectors of this process remains largely unknown. Using sequencing data from patient-derived xenograft samples and their matched primary tumors, we identified heat shock factor 1 (HSF1) as a critical regulator of this supportive transformation. This discovery of HSF1 within the microenvironment, led us to hypothesize that HSF1 was essential to drive this pro-lymphoma phenotype through stromal reprograming and extracellular matrix remodeling.

To address this, we utilized both in vitro and in vivo model systems to decipher the role of HSF1 in the most common subtype of non-Hodgkin’s lymphomas; Diffuse large B-cell lymphomas (DLBCL). To identify the potential mechanism with which HSF1 acted, we utilized an immuno-competent mouse model that had either wild type expression of HSF1 (wt), and a genetically engineered counterpart that was deficient in HSF1 for all cells throughout the animal (ko). Using a mouse HSF1-positive DLBCL cell line (A20), we implanted tumors in the flanks of both the wt and ko animals and tracked the growth kinetics of the tumors for 14 days. For wt animals, we observed the continuous growth of tumors over the 14 days, however, in the ko animals we observed a highly reproducible biphasic growth-shrinkage kinetics, with an inflection point at day 7.

We further investigated the lymphoma microarchitecture with a combination of microscopy and mass spectrometry, and tumor macro architecture using tissue mechanics assessment. Although at day 7, there were no significant difference in tumor sizes between wt and ko, we found significant differences in critical tumor properties, suggesting that the eventual shrinkage of the tumors was a secondary consequence of cascade of factors. Specifically, evaluation of tumor tissue sections revealed that the LME in ko mice exhibited lower cellularity, resulting in part from lower proliferation and increased apoptosis. Furthermore, these changes were accompanied by increases in intra-tumor adipose cells, mesenchymal stromal cells, as well as cytotoxic immune cells, and decreased suppressive immune populations. These findings combined with softer tissue mechanics, functional vascularization within tumors, increased porosity, collagen fiber density and alignment, led us to postulate that HSF1 deficiency within the stromal microenvironment induced a differential regulation of the tumor ECM composition and architecture.

To better understand the extent of this remodeling, we conducted a decellularization process to enrich for ECM proteins followed by mass spectrometry analysis for day 7 tumors derived from wt and ko animals, with spleens from both conditions used as controls. To identify relevant ECM proteins, we first compared our results to the database of ECM proteins within the Matrisome database, then identified the ECM proteins that were significantly enriched and depleted. We found that not only was the abundance of stroma ECM proteins greater within the ko tumors, but we identified differentially regulated ECM modules, some of which involved collagen functionalization by small leucine rich proteoglycans (e.g. Decorin, Biglycan). These results suggested that this differentially remodeled ECM in ko tumors may point to targetable vulnerabilities within the LME.

To evaluate whether we can achieve anti-lymphoma effects based on microenvironmental determinants found in ko LME, we took the list of significantly enriched proteins and identified testing candidates. Of note, we observed that the stroma-derived (mainly fibroblasts) proteoglycan Decorin was highly expressed in ko tumors. To demonstrate potential anti-tumor effects of Decorin, we implanted lymphomas into wt animals and allowed the tumor to establish for five days, after which animals were injected with Decorin for four consecutive days. From these experiments, we observed that injection of Decorin was able to significantly halt the growth of tumors. In addition, the introduction of wt fibroblasts into tumors growing in ko animals, and vice versa, demonstrated varying degrees of phenotypic rescue, suggesting that the stroma-derived ECM do play a role in inducing anti-lymphoma effects.

Together, we show that HSF1 is a critical molecular regulator of the LME, with deficiencies in HSF1 activity inducing an inability of stromal cells to support and sustain the growth of tumors in vivo. Furthermore, targeting the stromal microenvironment may be a viable strategy to induce vulnerabilities and anti-tumor effects within the tumor niche.