Assessing angiogenic responses induced by primary human prostate stromal cells in a three-dimensional fibrin matrix assay

Abstract

Accurate modeling of angiogenesis in vitro is essential for guiding the preclinical development of novel anti-angiogenic agents and treatment strategies. The formation of new blood vessels is a multifactorial and multi-stage process dependent upon paracrine factors produced by stromal cells in the local microenvironment. Mesenchymal stem cells (MSCs) are multipotent cells in adults that can be recruited to sites of inflammation and tissue damage where they aid in wound healing through regenerative, trophic, and immunomodulatory properties. Primary stromal cultures derived from human bone marrow, normal prostate, or prostate cancer tissue are highly enriched in MSCs and stromal progenitors. Using conditioned media from these primary cultures, a robust pro-angiogenic response was observed in a physiologically-relevant three-dimensional fibrin matrix assay. To evaluate the utility of this assay, the allosteric HDAC4 inhibitor tasquinimod and the anti-VEGF monoclonal antibody bevacizumab were used as model compounds with distinct mechanisms of action. While both agents had a profound inhibitory effect on endothelial sprouting, only bevacizumab induced significant regression of established vessels. Additionally, the pro-angiogenic properties of MSCs derived from prostate cancer patients provides further evidence that selective targeting of this population may be of therapeutic benefit.

Keywords: angiogenesis; mesenchymal stem cells (MSC); prostate; stroma; stromal paracrine factors.

Figure 1. Immunofluorescent characterization of the in vitro 3D fibrin matrix assay used to evaluate angiogenesis

HUVECs are allowed to adhere overnight to gelatin-coated dextran beads and embedded in a fibrin matrix to simulate the wound healing microenvironment. In the presence of conditioned media from stromal cells, endothelial tip cells invade the surrounding matrix followed by proliferating trunk cells; eventually forming a complex, branched, multicellular capillary network replete with patent lumens. This can be visualized by lectin (green) staining to identify the endothelial cells, which also produce a collagen IV-rich (red) basement membrane. Nuclei are stained blue with DAPI, indicating the multicellular nature of the vessels

Figure 2. Defining the critical stages and kinetics of angiogenesis in vitro

(A) Sprouting, the initial phase, involves tip cells probing and invading the surrounding fibrin matrix. (B) This is followed by the elongation phase in which trunk cells proliferate and migrate to form the growing vessel body behind the invading tip cell. (C) A complex, multicellular branched capillary network with patent lumens can then form via anastomosis of proximal vessels. (D) The dynamics of this process can be observed over a 10 day assay period with sprouting initiated by the addition of the stromal-derived factors. Once these sprouts begin to stabilize over the first few days, the elongation phase begins, continuing for the remainder of the assay. Lumen formation can be detected on approximately day 4–5 with branching and anastomosis of proximal vessels taking place over the final days of the assay. Also see Supplemental Movie S1.

Figure 3. Regression and regrowth of established vessels along pre-defined basement membrane tracts

(A) Regression of established vessels can be induced using an anti-VEGF treatment. (B) such as bevacizumab (10 μg/mL). (C) Regression of these vessels [lectin-positive, (green)] leaves behind an acellular [DAPI-negative, (blue)] basement membrane ‘tunnel’ defined by collagen IV (red) staining and indicated by the white arrowheads. Following drug wash out, tip cells indicated by white arrows can migrate along this basement membrane network to re-establish the pre-existing vessel network (D).

Figure 4. Primary human prostate stromal cultures enriched in MSCs and/or MPCs potently induce angiogenesis

(A) Conditioned media from primary stromal cultures derived from benign and malignant prostate tissue (nPrSCs and PrCSCs, respectively) induce robust vessel formation comparable to or surpassing that achieved using bone marrow-derived MSCs and normal human lung fibroblasts (NHLFs), the positive control in this assay. The uninduced negative control includes fully supplemented EGM2 media, but no stromal conditioned media was added. This response was quantified by measuring (B) cumulative vessel length and (C) the number of sprouts per bead. Importantly, conditioned media from prostate cancer cell lines (CWR22Rv1, LNCaP, and PC3) were also unable to induce sprouting in this assay (i.e. sprout length < 100 pixels, sprouts per bead < 1). All assays performed in triplicate. Ten beads analyzed per condition. All statistical comparisons made relative to NHLFs. *p < 0.05. **p < 0.005.

Figure 5. Targeting distinct phases of the angiogenic response using anti-angiogenic agents with distinct mechanisms of action

(A) The addition of conditioned media from NHLFs represents the positive control, while cultures containing fully supplemented EGM2 but no stromal-conditioned media represents the (B) negative control. Tasquinimod (10 μM), an allosteric inhibitor of HDAC4, significantly suppresses vessel sprouting when added upon assay initiation (C) but not when given on day 3 once the vessels have sprouted (E) Like tasquinimod, the anti-VEGF monoclonal antibody bevacizumab (10 μg/mL) profoundly inhibits sprouting when administered on day 0 (D) in addition to inducing a significant regression of established vessels when added on day 3 (F) as indicated by the presence of an acellular [lectin (green) and DAPI (blue) -negative] basement membrane sleeve defined by collagen IV (red) staining (G) Quantification of these responses was performed by measuring (H) cumulative vessel length and (I) the number of sprouts per bead. Assays performed in triplicate. Ten beads analyzed per condition. All statistical comparisons made relative to NHLFs. *p < 0.001. **p < 0.0001.