2017 Progress Report: Mammosphere Bioreactor For Life-Stage Specific Toxicology

EPA Grant Number: R835736C001
Subproject: this is subproject number 001 , established and managed by the Center Director under grant R835736
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).

Center: Vanderbilt Pittsburgh Resource for Organotypic Models for Predictive Toxicology
Center Director: Hutson, Michael Shane
Title: Mammosphere Bioreactor For Life-Stage Specific Toxicology
Investigators: McCawley, Lisa J. , Markov, Dmitry
Institution: University of Pittsburgh Main Campus , Vanderbilt University
EPA Project Officer: Klieforth, Barbara I
Project Period: December 1, 2014 through November 30, 2018 (Extended to November 30, 2019)
Project Period Covered by this Report: December 1, 2016 through November 30,2017
RFA: Organotypic Culture Models for Predictive Toxicology Center (2013) RFA Text |  Recipients Lists
Research Category: Safer Chemicals , Health , Human Health


Taking advantage of unique collaborations afforded through VPROMPT, we will develop a microfluidic Mammary Gland organotypic culture model (MG OCM) for toxicant assessment, for monitoring dynamic toxicant-dependent changes to cellular functions and signaling cascades, and elucidation of Adverse Outcome Pathways. In vivo the mammary gland is particularly sensitive to toxicant exposures during different life stages co-incident with times of active tissue growth and remodeling. We propose to extend our previously developed, microfabricated Mammary Gland Thick Tissue Bioreactor (MG-TTB) to support the combination of 3D culture methods with controlled exposure to investigate chemicals for potential mammary gland toxicity. We will utilize a simple mammosphere organotyic culture system and additionally evaluate a tri-culture system (normal human mammary epithelial cells + human mammary fibroblasts + sub-cutaneous adipocytes) for a more accurate recapitulation of mammary gland biology and crosstalk between heterotypic cells. We will validate this system for use as a medium throughput toxicology screening as well as for high information content analysis of targeted toxicant-dependent alterations to mammary formation. We will exploit this system for in vitro evaluation of potential environmental toxicants for effects on mammary development using both (1) chronic exposure and (2) acute exposures co-incident with various stages of gland development. Furthermore, this system will be used for high content analysis of putative toxicants with a focus on key biomarkers. Our specific objectives to validate the microfluidic OCM for toxicant are as follows: (A) to develop and validate the predictive utility of self-contained, fully-automated MG OCM modules; (B) to investigate the effects of chemical exposures on MG OCMs for chemicals shown to reduce lactation index in only the F2 generation of multigenerational studies – strongly suggesting a role for life-stage specific exposure; and (C) to develop and validate toxicant assessment for compounds requiring metabolic activation using a paired Liver-OCM/MG-OCM.

Progress Summary:

Mammary Gland Triculture. We began studies this year evaluating media compatibilities for a tri- culture system incorporating normal human mammary epithelial cells, human mammary fibroblasts and sub-cutaneous adipocytes (ScienCell) to determine a single common media composition that would provide adequate growth signals to these varied cell types. We observed that pre-adipocytes cultured in Maintenance (PAM) versus differentiation Media (PADM) yielded morphologies consistent with differentiation such as larger flatter cells with oil redO staining of cellular compartment. The DMEM/F12 Complete media appears to induce proliferation and differentiation of the pre-adipocyte cultures. Pre-adipocytes did poorly in DMEM Complete supplemented with DAdDS. In parallel, MCF-10A and MCF-12A were cultured in 3D ECM in DMEM/F12 Complete and DMEM/F12 Complete media supplemented with PAdDS to assess mammosphere formation in the presence of PAdDS. Addition of PAdDS let to suboptimal mammosphere growth and resulted in complete cell death following 7 days in culture. Thus, due to the issue with both cell types to DMEM/F12 Complete+PAdDS, we are currently testing maturation of pre-adipocytes in 3D gelled ECM with a focus on using DMEM/F12 complete as our common media.

Toxicant Effects on Mammary Formation. Our mammary gland (MG) bioreactor platform is standardized to a 96-wellplate footprint, which enables us to conduct image-based screening with instrumentation such as the Molecular Devices ImageXpress Micro XL. In addition, we have made bioreactor modifications to (1) allow for distribution and overlay of matrix containing fluorogenic proteolytic sensors of established mammospheres in the MG OCM system and (2) establish base- line sensitivities of proteolytic biosensors and protocols. Alongside our bioreactor development, we have tested conditions and sensitivity of proteolytic probes focusing on (1) determining baseline proteolytic molar range of sensitivity of cleaved versus uncleaved proteinase probes (i.e., DQ-ECM and a series of custom in house generated proteolytic beacons (PB)); and (2) determining the matrix compositions that yield the least background autofluorescence on the HTS ImageXpress microscope. To generate a more high-throughput screening method to test these proteolytic probes, we made a custom 96-well stamp that would help to constrain the spheroid culture volume thus reducing a number of imaging sites in x-y plane as well as the size of z-stack for each site (Figure P1-2A, B). Once gelled, cells were plated on top of this layer and cultured in media containing 2% matrigel, conditions that permit mammosphere formation (often referred as 2.5D culture).

Figure P1-2. (A) 96-post array compatible with 96-well plate for hydrogel molding to create a flat area in the center of the wells promoting spheroid formation w/ a confined volume, which simplifies imaging. 1 – post array; 2 – hydrogel molding with the post array; 3 – bottom view of the flat area molded in hydrogel. (B) Schematic view of the individual well showing molding and spheroid grown overlaid with beacon containing ECM. (C) Schematic diagram of the fluorescent proteolytic beacon that contains non-cleavable red dyes (reference) and green fluorescent reporter dyes attached to the molecule via a variety of cleavable linkers (1 - gelatinase, 2 and 3 - collagenase).

Mammospheres were then overlaid with hydrogels containing Proteolytic agents (Lower Panel, Figure P1-2B). We tested a variety of DQ-ECM reagents including DQ-Gelatin; DQ-ColI; DQ-ColIV and DQ-BSA or with custom proteolytic beacons that have the sensitivities to gelatinase and collagenase activities as indicated in Figure P1-2C.

Our findings to date indicate that DQ-Gelatin fluorescence from proteolytic activity of the cells is hard to see above the Matrigel autofluorescence. Activity measurements with PBs rely on a two- color radiometric approach where non-cleavable red fluorescing dye provides the reference signal and cleavable green fluorescing dye indicates proteolytic activity (Figure P1-3A). Results with proteases probes imaged in matrix (Matrigel and/or Matrigel/hydrogel) using the HTS microscope system and in 96-well plates have been obtained (Figure P1-3).

Figure P1-3. Imaging mammosphere-associated MMP activity from z-stack data. For each x-y position within the well both white-light (WL, phase-contrast) and fluorescence (red, REFERENCE TMR and green, SENSOR FL channels) images were collected at 31 focal planes (z-position, 15 μm apart) to capture positions above and below the upper/lower gel interface. (A) Enlarged views of focal planes z12 and z17 show WL images and corresponding red and green fluorescence images for beacon 1, cleavable by gelatinases. Increased green fluorescence indicates activity. (B and C) Graphs show typical fluorescence data, quantified from appropriate regions of interest, as a function of focal-plane for both organoid-associated fluorescence. (B) Organoid #1 sensor (signal/noise, S/N =15), reference (S/N = 11) and S/R ratio (S/N = 9). (C) the background (control) fluorescence and S/R ratios for control versus organoids, #1, #2 and #3, with maximal S/R ratios (mean ± SD) of 0.18 (±0.02), 0.45 (±0.05), 0.26 (±0.03) and 0.34 (±0.04), respectively, and z-stack-averaged S/R values (mean ± SD) of 0.17 ( ± 0.01), 0.39 ( ±0.06), 0.24 ( ±0.01) and 0.33 (±0.02), respectively, showing significant organoid-associated gelatinase-activity for each imaged mammosphere.

Figure P1-4. MCF-10A Mammosphere Growth Overtime with Increasing Log Dose of Toxicant Genestein. (A) BF images of MCF-10A mammospheres from Day 10 and Day17 of culture with increasing log dose of Genestein. (B) Average MCF-10A mammosphere area over time with increasing log dose of genestein (1, 10, 100 and 1000 nM) introduced on day 3 of culture.

Findings to-date with PBs indicate that: (1) the beacon is retained in sample matrix for at least 96 hours after overlay; (2) two-color beacons imaged on ImageXpress microscope detect significant proteinase activity associated with established organoids; (3) preliminary analysis indicates more robust response with the gelatinase versus collagenase beacons (i.e., PB beacon 1 versus PB beacons 2 or 3 in Figure P1-3C); (4) initial results suggest differential responses with organoids with MCF10A > MCF12A > MCF7; (5) the HTS imaging approach allows scoring of organoid-associated beacon activity for a significant number of organoids (6 fields/well, 2-5 organoids/field) over time (up to at least 96 h); and (6) overall, known toxicants can now be tested using MCF10A & MCF12A organoids with the gelatinase beacon to evaluate sensitivity and specificity of the assay.

Pilot testing of mammosphere formation in the absence and presence of 5 log dose application of toxicant (Bisphenol A, Genestein and Nonylphenol) is shown in Figures P1-4 through 6. We used our dual labeled MCF-10A and MCF-12A cells lines with stable expression of both lentiviral sensors pLVXIRES-puro-FusionRed-H2B (Histone 2B) and pLVX-puro-casper-3BG (Caspase-3 activity sensor). Cells were challenged on Day 3 of culture to control for even cell seeding.

Overall, we can detect changes in mammosphere growth when challenged with all three compounds within ranges of sensitivity we expected based upon known bioactivities. There is some difference in timing between the cell lines with Nonylphenol challenge of MCF-12A generating changes in mammosphere size at time points slightly earlier in culture than does the MCF-10A line (Figure P1-6 and Figure P1-5 for MCF-12A and MCF-10A challenge results respectively). Overall, these studies are providing baseline values as we begin the assessment of the larger toxicant panel in the MG-OCM.

Figure P1-5. MCF-10A Mammosphere Growth Overtime with Increasing Log Dose of Toxicants Bisphenol-A (A) and Nonylphenol (B). Avg. MCF-10A mammosphere area over time with increasing log dose of Bisphenol-A (A) or Nonylphenol (B) at 1, 10, 100 and 1,000 nM introduced on Day 3 of culture.

Figure P1-6. MCF-12A Mammosphere Growth Overtime with Increasing Log Dose of Toxicant Nonylphenol. Avg. MCF-12A mammosphere area over time with increasing log dose of Nonylphenol at 1, 10, 100 and 1,000 nM introduced on Day 3 of culture.

To provide varying toxicant concentrations to individual cell culture chambers within our bio-reactor cartridge we integrated a commercial 6-position 7-port selector valve, which allows us to utilize a single micropump to perfuse six chambers via individually applying a time-division multiplexing approach. We are working closely with Project 5 on hardware integration.

Liver-MG Integration. To achieve functional interface between Mammary OCM and the Liver Module, we focused on developing common media culture conditions. Short-term (4-Day) culture testing of viability of cell monolayers demonstrated that liver-conditioned culture media could maintain mammary epithelial cell lines. Morphogenesis tests of mammary epithelial cell lines over time revealed that the morphogenic process was blunted when cells were directly cultured in liver-conditioned culture media; however, a 50/50 mix of liver-conditioned and naïve DMEM/F12 complete media resulted in mammosphere formation comparable to that established by DMEM/F12 alone. In Year 4 we will evaluate mammary formation in the presence of conditioned media from the liver organoid system that was exposed to a toxicant. Intense dialogue between Projects 1, 4 and 5 resulted in the strategy for Liver – MG OCM integration shown in Figure P1-7.

PDMS Chemical Interaction Testing. We have completed our assessment of chemical interactions with polydimethylsiloxane (PDMS). Measurements were done at 48, 96, and 216 h of incubation via UV-Vis absorption. We have found that for the compounds that would adsorb to PDMS surface majority of such interaction was complete within the first 48 h. Although there were no strong correlations among any one chemical characteristic there is a correlative relationship with H-bond donor number and the octanol-water partition coefficient (Log P) to % absorbed. There appears to be a threshold for Log P <1.27—1.83 where chemicals are not adsorbed by PDMS. However, not all chemicals above this threshold are adsorbed either.

Figure P1-14. Schematic diagram of the Liver – MG OCM integration approach. Based on the flow rate compatibilities one Liver OCM can supply 3 chambers of the MG OCM. Since mammospheres do not grow well in Liver effluent alone Reservoirs R3 and R4 will act as fluidic buffers while valves V1 and V2 will allow to sequentially withdraw fluids from either R3 / R4 or R5 with fresh mammary media thus creating a 50:50 mix. Selector valve V3 will allow MG Pump to sequentially draw mixed media through each of the culture chambers C1 - C6.

Future Activities:

  • We will continue collaboration with Project 5 on upgrading MG OCM to include customized valves and pumps necessary for media mixing, sampling, and coupling the liver-OCM upstream of the MG-OCM.
  • We will generate a mathematical description of toxicant/bioreactor interaction.
  • Screen toxicants: validate MG OCM reactor system for medium through-put for stage specific and acute toxicity assessment using optical and electrochemical sensing.
  • Move towards High Information Content profile analysis beginning with proteinase activity analysis of mammosphere formation using identified toxicant.
  • Validate strategies for Liver-MG-OCM coupling.

Journal Articles on this Report : 1 Displayed | Download in RIS Format

Other subproject views: All 14 publications 3 publications in selected types All 3 journal articles
Other center views: All 149 publications 39 publications in selected types All 39 journal articles
Type Citation Sub Project Document Sources
Journal Article Karolak A, Markov DA, McCawley LJ, Rejniak KA. Towards personalized computational oncology: from spatial models of tumour spheroids, to organoids, to tissues. Journal of the Royal Society Interface 2018;15(138):20170703. R835736 (2018)
R835736C001 (2017)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: The Royal Society-Full Text HTML
  • Abstract: The Royal Society-Abstract
  • Other: The Royal Society-Full Text PDF
  • Supplemental Keywords:

    Mammary development, mammary toxicology, organs on chip, mammary on chip, PDMS bioreactors, thick tissue bioreactor, PDMS interactions, organotypic cell culture

    Relevant Websites:

    Vanderbilt-Pittsburgh Resource for Organotypic Models for Predictive Toxicology (VPROMPT) Exit

    Progress and Final Reports:

    Original Abstract
  • 2015 Progress Report
  • 2016 Progress Report
  • 2018 Progress Report

  • Main Center Abstract and Reports:

    R835736    Vanderbilt Pittsburgh Resource for Organotypic Models for Predictive Toxicology

    Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
    R835736C001 Mammosphere Bioreactor For Life-Stage Specific Toxicology
    R835736C002 Organotypic Culture Model to Analyze Developmental LimbMalformationsResulting from Toxicant/Teratogen Exposure
    R835736C003 Validating a fetal membrane on a chip model for characterizing reproductive toxicant exposure risks
    R835736C004 Organotypic Liver Model for Predictive Human Toxicology and Metabolism
    R835736C005 Systems Engineering & Analysis for Organotypic Culture Models