(128a) High-Throughput DNA-Directed Patterning As a Tool to Study Prostate Cancer-Bone Marrow Niche Interactions at Single Cell and System Level | AIChE

(128a) High-Throughput DNA-Directed Patterning As a Tool to Study Prostate Cancer-Bone Marrow Niche Interactions at Single Cell and System Level


Kozminsky, M. - Presenter, University of California, Berkeley
Sohn, L. L., University of California at Berkeley

Transforming cancer from a terminal diagnosis to a chronic or curable disease rests on predicting and pre-empting disease trajectory. This is particularly true in prostate cancer, in which a disseminated tumor cell (DTC) may reside in the bone marrow and remain dormant for years, only to awaken into metastatic cancer [1]. While DTCs are detected in the bone marrow of many prostate cancer patients, only 1% of these cells yield macrometastases [2]. Thus, there is a pressing need for a model to study the characteristics of the interactions within the microenvironment that influence prostate cancer progression. Animal models do not accurately recapitulate human disease progression [3], while current in vitro systems lack the necessary fine-tuned control and throughput to pursue the underlying mechanisms of dormancy. To increase the resolution and scale of in vitro systems, we used high-throughput DNA-directed cell patterning to fabricate multiple replicates of an in vitro bone marrow microenvironment to study prostate cancer tumor-cell dormancy. By harnessing the power of this innovative patterning method, high-throughput single-cell studies that range from probing the individual contributions of different microenvironmental cell types to creating in vitro microenvironments of increasing complexity can be performed to study the conditions that lead prostate cancer cells to remain dormant or to proliferate.


High-throughput DNA-directed single-cell patterning was used to create replicates of an in vitro bone-marrow microenvironment within an easy-to-image conventional chamber slide (Figure 1a). Briefly, positive photoresist was spin-coated onto an aldehyde-functionalized glass slide and exposed to UV light through a mask [4]. Amine-terminated 20 nucleotide oligos were dropcast onto the patterned slide and conjugated via reductive amination. Photoresist was then stripped. This process could be repeated to conjugate different nucleotide sequences to different regions of the slide. For the microenvironments used in this work, up to four different iterations of patterning were performed, corresponding to up to four different oligonucleotide sequences. Finally, a layer of polyacrylamide was patterned to confine cells to specific areas (“microislands”). Cells were then tagged with the complimentary oligo [5] and introduced to DNA-patterned regions. Hybridization of complementary oligos subsequently followed, leading to the specific patterning of cells. For single-cell interaction studies, the configuration of the photomask allowed for the fabrication of four arrays of 1250 microislands (or a total of 5000 microislands), each containing one prostate cancer cell and one microenvironment cell, that were enclosed within a four-well chamber slide. The distance between cells within the microislands ranged from 30-296 µm. For the high-complexity microenvironment studies, the configuration of the photomask allowed for the fabrication of four arrays of 81 microenvironments (or a total of 324 microenvironments) that were enclosed within a four-well chamber slide. In addition to prostate cancer cells (represented by the cell line PC-3), microenvironment cells included osteoblasts (represented by MC3T3), osteocytes (represented by OCY454), macrophages/pre-osteoclasts (represented by RAW264.7), and endothelial cells (represented by HUVEC).


Performance of the patterning process was verified with complimentary oligos tagged with fluorophores and with cells labeled with CellTracker dye. Initially, cell types selected for inclusion were cultured with three media conditions (alpha-MEM, DMEM, F12K) to select the best media for co-culture. Alpha-MEM maximized viability across cell lines.

To investigate individual contributions of the different bone cell types found to influence the proliferative or dormant tendencies of prostate cancer cells, PC-3 was patterned in individual co-cultures with either HUVEC, MC3T3, or RAW264.7 (Figure 1b). Proliferation rate and morphologies were assessed over the course of seven days. Preliminary data showed slower growth rates of PC-3 (i.e. higher doubling times, ranging from 50-77 hrs) in the presence of microenvironment cells than the reported values, as low as 25 hrs, of these cells, alone [6]. RNAseq was also used to examine cellular changes based on interactions.

An in vitro niche was then fabricated and validated for the ultimate use of examining the broader system of interactions between the cells of the bone marrow and prostate cancer cells (Figure 1c). Cellular phenotypes and behavior were verified using RNAscope (in situ hybridization) and immunofluorescence staining. Mineralization by bone cells was verified by Alizarin Red S staining, while osteoclast differentiation was visualized with tartrate-resistant acid phosphatase (TRAP) activity staining. Finally, low numbers of fluorescently-labeled PC-3 cells were patterned along with bone cell types and imaged over the course of several days to assess proliferation, thereby determining the microenvironment influence on these cells. The highly proliferative PC-3 cells showed overall low proliferation in the presence of microenvironmental cells.


Applying high-throughput DNA-directed cell patterning system yielded both high replicate single cell co-cultures and a highly complex cellular model system with unprecedented control, scale, and throughput. Future directions include exploration of interactions using primary cells with long-term goals of incorporation in personalized medicine. These systems can be applied to answer fundamental biological questions about prostate cancer proliferation rates and dormancy in secondary locations.


[1] Morrissey C et al Journal of Molecular Medicine 2016 Mar 94 3 259 65

[2] McAllister, S. S. & Weinberg, R. A. Nature cell biology 16, 717–727 (2014).

[3] Bragado, P., Sosa, M. S., Keely, P., Condeelis, J. & Aguirre-Ghiso, J. A. in Minimal Residual Disease and Circulating Tumor Cells in Breast Cancer (eds. Ignatiadis, M., Sotiriou, C. & Pantel, K.) 195, 25–39 (Springer Berlin Heidelberg, 2012).

[4] Scheideler, O. J., Yang, C., Kozminsky, M., et al (2020). Recapitulating complex biological signaling environments using a multiplexed, DNA-patterning approach. Science advances, 6(12), eaay5696.

[5] Todhunter, M. E. et al. in Current Protocols in Chemical Biology (John Wiley & Sons, Inc., 2016).

[6] Liu J et al. Oncol Lett. 2020 Nov;20(5):230. doi: 10.3892/ol.2020.12093.