The liver, brain, and bone tend to be ?hot spots? for metastatic colonization. However, what determines where specific tumor cells metastasize within the same patient and across patient populations remains poorly understood.

Created by Sara Gosline sgosline
Hi Aedin @aedinc we can look at cell stiffness directly...to complement/validate gene expression if it is relevant signature.
Hi I am a computational biologist. So I can contribute analyses of TCGA, GTeX, CCLE and any of your data. I am very excited to work on this. I am still mulling the mechanism of dissemination (blood, lymph)... But on cold/hot sites. I started extracting gene signatures of tumor cells exposed to different ligands or cytokines. The idea being that the transformed tumor microenvironment may provide molecules essential for the tumor. The signature should not be highly expressed by the normal non-malignant site. Using these signatures I have identified normal tissues with higher levels of the signature (AUC/ROC analysis) using the GTeX normal tissue atlas data. I have 1 signature that is not expressed in normal breast but is expressed by transformed fibroblasts, is expressed in ER+ breast tumors (bulk tissue) which have a strong propensity to bone and is highly expressed and specific to bone marrow. This signature is really low in other tissues.. so we may have a bone met signature for ER+ breast cancer ... this analysis came together very easily and it maybe a nice pipeline that we could apply systematically to address some of these questions. I have started looking at essentially genes (CRISPR) in different tumor cells lines, and this receptor is an essential gene in ER+ Breast cell lines. So perhaps we could make a list of all of the ligands that might be secreted by transformed microenvironment (that might be "additive" for the tumor). I can cross-check the list of receptors expressed by tumors against CRISPR/achilles to refine the list of signaling molecules that the tumor "needs". I can use existing public data, or someone in this collaboration could expose tumor cells to the ligands to generate profiles, gene signatures of response to ligands. Then using GTeX I will identify tissues that express these signatures. In parallel, I can make a database of genes that are highly specific to sites with high met propensity, low met propensity etc. I can integrate this with TGCA, ICGC, Metabric etc etc Alternatively, we could look at cell-stiffness, via looking at gene expression of ECM genes. I think we could tie this to evolution Aedin
@milesm we can all meet at 4pm then
Hi especially to Luis and Eleonora (who had signed up for the Fri. 1pm slot)- I can no longer make 1pm tomorrow, but will try to tune in at 4pm (looks like you will be on that call anyways). I look forward to "meeting" everyone on the phone. Thanks - Miles
Hi Eleonora, intravital imaging would be a great asset, and very informative tool!
@sole sure! looking forward to discussing some interesting ideas
Hi Eleonora, @EDondossola- this is really interesting. Maybe, if you are available, you can join us this Friday at 4 pm to discuss ideas. It will be nice to have your expertise in intravital imaging to address the question on how microenvironment affect dormancy and survival of DCCs in these 3D printed models.
@soloriol, hello I find your system very interesting. I have experience in imaging biomaterials implanted in the skin and I can provide intravital imaging support for your 3D printed microenvironments. With this approach we can complement more conventional tecnhiques and monitor the dynamics of the process and the interplays between cancer cell and stroma (ECM and cells, e.g. immune cells) overtime, up to weeks.
Friday at 1:00?
Hi all -- I propose a teleconference on Friday. All three remaining time slots (9, 12, 1) work for me. If anyone has a preference, maybe go ahead and sign up for a particular time slot and I'll add my name to it. Sound ok?
Happy to join in a teleconference here. Did we arrange a time?
@soloriol Got it!! makes sense - good night!
I'm a bit tired, so I definitely wasn't clear. I was meaning evolution post dissemination.
@soloriol this seems like a great idea, but given the need of targeting metastasis I wonder if the same topic and approach would better serve the focus by modeling it in the context of the target organ rather than the primary tumor, as you mention in your last comment. Unless you mean evolution of primary tumor cells post dissemination?
I would love to try the different dormancy models on the microenvironments. I really like the idea of trying to mimic the evolution of the primary tumor. I've really gotten interested in the idea of immune mediated dormancy, and would like to see if we can create a system that would allow for the immune cell dominance vs cancer cell dominance and see if we can elucidate how to shift the balance.
Hi Luis, @soloriol, I am really interested in your bone- and lung-like environments to test the contribution of niches to the quiescence of DCCs within the same organ based on secreted factors. We have studied for several years now which factors determine a more restrictive (cold) or more permissive (hot) microenvironment for DCCs. We could convert a restrictive niche into a permissive one and vice versa and interrogate their ECM and stiffness. Our group also has nice dormancy models/cell lines to explore this from single cell to clusters and micro/macromets level. It will be also nice to see the contribution of immune cells to the quiescence and latent stemness of DCCs. We have preliminary data on this topic that maybe we can explore. Even more, we could explore how those DCCs that left primary sites before a tumor was detected persist in a dormancy state for almost the whole evolution of the primary tumor. We could ask is there any signature that make them evade immune system? or do they create a different ECM composition that allows for an extended quiescence period? Best, -sole
Hi Everyone, I'm a bit of a lurker and not well versed in message boarding, but this is a really interesting topic. My group has recently developed a 3D printing system that allows for the development of modular ECM based microenvironments. We have been able to get microenvironments using fibronectin in the fibrillar conformation similar to what is observed in the tissue, laminin in both a more basal lamina-like conformation as well as in a fibrillar conformation, collagen, as well as different ECM mixtures that contained tenascin-c. We have had good luck enriching the cancer cell population of pleural effusions, from patients with breast cancer in the fibronectin based environments. We have also used the system to make a bone-like environments that were the site of metastasis after intracardiac injection of breast cancer cells. It might provide a nice system for really interrogating the role of the ECM.
Hi @milesm, @zawang, @jungwoo, @aedinc, @jjbravo, @sole, @subhode, @alexandra.naba, and @lanilonzo, There might be a number of topics here that could become a project!   To move it along further I suggest signing up for a teleconference this week. The signup sheet can be found [online](https://docs.google.com/spreadsheets/d/1hY53jRaqoBMnb9HhuE8dN3k4ejX37gG-4C4pattxkuM/edit#gid=1701101959) and as many of you can sign up for as many topics as you'd like. Just put the 'call topic' in column F. The workshop organizing committee is ready to host and facilitate these calls as need be, all you need to to do is call in the number in the column A. You can use this forum or communicate offline to find a time when all of you can call in to discuss further.   The sooner you circle around a project the more time you will have to prepare for a successful project in June. Please feel free to ask me if you have any questions about the process.   -sara
Hi all -- joining in this already well-rounded conversation, I also would like to suggest looking at fluid flow and vessel architecture in "hot" and "cold" sites. I imagine the properties of tissue perfusion, fluid pressures, and vessel permeability would provide complimentary perspective to the many approaches and focuses (especially ECM, hypoxia , and stiffness) already mentioned above. As a technique, I perform intravital imaging to measure heterogeneous vascular behaviors, especially as they relate to infiltrating immune cell populations. There has been recent work discussing the role of exosomes and microvesicles "priming" the pre-metastatic niche (e.g., Hoshino et al., Nature 2015), and I would also be interested to know how ECM/stiffness/vascularization influence the organotropic deposition of nanoparticles, exosomes, and microvesicles. Has this thread scheduled a teleconference this week? Thanks -
@jungwoo Our system is in vitro. will be good to pair it with in vivo system. our focus is on understanding why some tissue can harbor metastasis and how the environment will affect treatment.
I think you need to incorporate a temporal variable into your "hot" and "cold" sites. Because frequency of metastasis does not mean that they share the same temporal time line for development of metastasis. In addition, this hot and cold sites vary with the cancers. For example HNSCC and gastric cancers give very rarely bone mets, but DTCs in bone marrow are present at a higher frequency than the mets develop suggesting that the organ is in many cases "cold". That the cells are rare in some cases is just how it is and that is why we need new modalities to innovate in how to study rare cells. This group has discussed quite a bit, it may be a good time to define aims or questions and a set of approaches to have ready for the teleconferences discussions.
Hi @zawang, I have a very similar question what you have - what is the general microenvironmental features between "hot" and "cold" metastatic niches. The major challenge in conventional animal models is that disseminated tumor cells in "cold" metastatic tissues are extremely rare and difficult to study downstream cellular processes. Are your models in vitro or in vivo? My group develops in vivo metastatic niche models based on tissue engineering approach, which can be accessible under an intravital microscope and also increase a chance of metastatic events. While we have been focused on humanized mouse models, this can be easily changed to a syngeneic system and also in vitro or ex vivo system. @jjbravo has excellent systems and techniques to ask this question. It would be fantastic to ask this question together with engineered metastatic models.
we have unpublished data--still working with the editor to get through what will be the second revision
Hi @zawang very interested point! I will be happy to help with the biology! Are the data regarding ECM/decellularized tissue published to look at them?
also, want to point out that "bone mets" are really bone marrow metastases. they start in the bone marrow, which is why most of the bone metastasis happen in spine, pelvis, etc and in the middle of the bone---areas with bone marrow. This is also we almost never see hand, feet mets or metastases at a joint
The "hot" metastatic niches are lung, liver, bone marrow, lymph nodes and brain (least common). I think the bone/bone marrow and lymphatic mets have to do with cell-cell interactions. My group has data on the liver and lung ECM/decellularized tissue affect metastasis. Brain is difficult for us to model as it has very little ECM, we tried. To answer an earlier question, we have compared the mass spec data between liver and lung, they are quite different. The most interesting aspect of the organ specificity is not only where tumor cells can establish themselves, but also why they can't live in other highly vascular organs--most of the GI tract, kidneys, spleen, skin, muscle, etc. We have the model to study this but need biologists to help with some of the pathway work
Hi @alexandra.naba Thanks for your reply! As @sole point out we can have tumor co-stain for specific micro-environment features or use tumors expressing specific pathways biosensors ( such as hipoxic sensor) to identify specific regions. Blood vessel density, macrphage presence or absence can also be used to identify specific microenvironments with specific motility patters. In this paper ( http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3908591&tool=pmcentrez&rendertype=abstract) we show that vascularized and non vascularized microenvironments are indicators of single vs streaming types of motility in vivo by intravital imaging. While areas that are non vascularized seems to have more single cell ameboid movement we dont know for example, what are the characteristics of ECM in those regions vs more vascularized areas. Javier
Hi @alexandra.naba, maybe worth reading this paper (https://www.ncbi.nlm.nih.gov/pubmed/28114271) regarding your second question. Here the authors identified that hypoxic tumor regions released DTCs that when arrive to secondary organs were more prone to become dormant and evade therapy. Maybe the staining of primary tumor with pimonidazole adducts may reveal those regions and you could do LCM on those. Still, don't know the technical difficulties of doing profiling of ECM.
@JAguirre-Ghiso I like very much the idea of interrogating the ECM (composition, mechanics) and tumor-cell response (@jjbravo, this is your expertise) to the ECM when single DTCs transition to form small cell clusters and then move on to form micro and macromets. We could within the certainly do it in the same organ or even in bone marrow vs lung. And if we worked with xenografts, we could also look at tumor-derived vs stromal-derived signals. The proteomic profiling of the ECM surrounding DTCs may be technically challenging also I could think of a couple of technical modifications we could implement to our pipeline to do it. @jjbravo to answer you rquestions: 1) yes we can!; 2) certainly, but this is more challenging: we have previously published the comparison of the ECM of different region of the same colorectal tumors and found an almost complete overlap (Naba et al, BMC Cancer, 2014). So we could refine this analysis with a focus on regions of tumors that are giving rise to DTCs, but I would not know how to identify these... Thoughts?
@alexandra.naba Can we extract ECM from different regions of an organ and do proteomics analysis ? I think looking at tumor cell phenotype ( motility, invasion) is a great idea since we know that the modes of migration are highly dependent on matrix composition and stiffness. We can look, in the different panels of tumors based on their ECM composition, what migration modes are more predominant and where DTC seed. How do these DTC will behave in the metastatic organs and how they remodel the ECM over time? We can combine your proteomics analysis with intravital imaging and imaging windows to extract those dynamics features. @alexandra.naba shouldn't we consider also that within the primary tumor we may have pockets of different extracellular matrix and stiffness that may give arise to specific DTC populations? The tumor may have a particular ECM signature but the presence of specific microenvironments with different composition may prone DTC in those niches with a particular feature, increasing the complexity and the heterogeneity of the tumor. Have a great day! JAvier
@alexandra.naba. Alexandra, this is a very pertinent observation. In our studies (this may be relevant to @scarc Giuliano) we find that initially single DTCs and small cell clusters are under control of the cues in the microenvironment that drive their initial dormant phenotype. I the bone marrow but recently also in the lung we found that TGFb2 is an inducer of dormancy. In the bone marrow there is more TGFb2 overall than in the lungs (https://www.ncbi.nlm.nih.gov/pubmed/24161934). Still we find niches rich in TGFb2 in the lung but even in the same lung lobule a few cell diameters away a growing metastasis has managed to turn off all TGFB2 signal (https://www.ncbi.nlm.nih.gov/pubmed/28114271). So I agree that the metastasis once it reaches a critical mass that can reproduce the ECM and environment of its tissue of origin overrides the organ growth control signals. The transition of single DTCs, to small cell clusters, to micro and macromets may be a good model within the same organ to test how ECM composition and stiffness may also instruct the stroma to stop making growth suppressive cues. You guys may be awn a good path to start defining specific questions.
This recent and excellent review from Yibing Kang's group may be a very useful resource for this topic. Determinants of Organotropic Metastasis. (2016) HA Smith, Y Kang. Annual Review of Cancer Biology. http://www.annualreviews.org/doi/pdf/10.1146/annurev-cancerbio-041916-064715 The spread of cancer from a primary tumor to distant organ sites is the most devastating aspect of malignancy. Dissemination to specific organs depends upon blood flow patterns and characteristics of the distant organ environment, such as the vascular architecture, stromal cell content, and the biochemical milieu of growth factors, signaling molecules, and metabolic substrates, which can be permissive or antagonistic to metastatic colonization. Metastatic tumor cells possess intrinsic cellular properties selected for adaptation to specific organ environments, where they co-opt growth and survival signals, undergo metabolic reprogramming, and subvert resident stromal cell activities to promote extravasation, immune evasion, angiogenesis, and overt metastatic growth. Recent work and new experimental models of metastatic organotropism are uncovering crucial details of how malignant cells metastasize to specific tissues, revealing key mediators that prepare metastatic niches in specific organs and identifying new targets that offer attractive options for therapeutic intervention.
Hi all, I am joining the conversation as I am very interested in identifying microenvironmental regulators of metastatic tropism. The biomechanical properties of the tumor microenvironment are in part governed by the protein composition of the extracellular matrix of the tumor microenvironment, which is what I study. What we have shown using proteomics is that the ECM composition of sites to which tumor cells metastasize (for example lung or liver) does not necessarily resemble the ECM of primary tumors from which metastases derive from. However, the ECM of metastases resemble the one of the primary tumors they originate from. So we can hypothesize that the tumor cells, when they reach a distant organ, produce their own ECM and that has an impact on the stiffness/elasticity of the host tissue. I think it would be really exciting to take a panel of primary/secondary tumors of different origins and that metastasize to the same or different organs and look at 1) their ECM using biochemical and biophysical approaches, and 2) look in parallel at the tumor cell phenotypes (Javier?). What do you think?
I've heard some talk about epigenetics being hypothesized as a critical mechanism for this. Is this something someone in the team is looking into or there is data available for? Has that been associated with stiffness?
@aedinc This paper contains measurements of lymph node stiffness with and without mets and shows: 1- how stiffness changes when the LN has mets and 2- how stiffness of LN predicts mets. http://onlinelibrary.wiley.com/doi/10.1002/(SICI)1097-0142(19971115)80:10%3C1920::AID-CNCR8%3E3.0.CO;2-R/full This is a review on stiffness of different tissues: http://www.tandfonline.com/doi/full/10.1080/15476278.2015.1019687 Another one showing how stiffness of mammary primary tumor correlate with mets. They also measure stiffness of tumors compared and with normal mammary fat pad: https://www.nature.com/articles/srep05512 This is study also show a mechanism for how matrix stiffness can drive metastasis: https://www.nature.com/ncb/journal/v17/n5/full/ncb3157.html Hope this helps! I will keep looking in my archives to see if I found more. Have a great day! Javier
Interesting. Can you point me to references that show differences in stiffness of different Mets. Also stiffness of different tissues. Maybe I can correlate these back to GTex, TCGA etc data and compare the gene signatures.
Indeed very interesting topic. Are there studies or evidence for the relevance of physical cues of the microenvironment in the heterogeneity of metastatic colonization? For example, we know that bone and brain have dramatically different stiffness. And several fundamental studies have shown enhanced growth/proliferation, enhanced motility of cells when they meet the "right" microenvironment stiffness.
@aedinc following the conversation I found this study in where they identified markers relevant to lymphatic interaction by screening peptide libraries. Here it is: http://www.pnas.org/content/109/47/19280.long Javier
This is a very interesting topic @aedinc Aedin! We don?t completely understand how tumor cells disseminate through the lymphatic system. Do they use the same mechanisms as for blood vessels intravasation, meaning actin-rich invadopodial protrusions? Can we find cells in the TME that are prone for lymphatic spread? What is its genetic bckg? How treatment potentiate can one or the other (as the paper you mention?) Can we model this process in vitro? Aedin, maybe profiling the DTCs in the lymph node can shed light into the molecular pathways driving this process, and tell us about specific primary tumor microenvironments that select for those cells. Do you know if there any study on this? Looking forward for this exciting workshop! Javier
Aedin, thanks for getting the discussion started. Regarding your question about sharing references, some options are to either share hyperlinks directly in the discussion thread, or you can create a shared references folder in the IACM project workspace located here here: https://www.synapse.org/#!Synapse:syn8063299. That folder can be used as a shared space to upload relevant publications. Let us know if neither of these options are suitable and we can look into other mechanisms such as the one you suggested. There might be other functionalities within Synapse that I am not aware of, and Sara can comment on that. Nas
Different tumors have preference for different secondary sites, but different tumors may also have preference for lymphatic or hematogenous dissemination. There have been recent publications to support the latter in breast and colon cancer. The complexity of metastases and potential importance of lymph involvement was recently highlighted by Dufies et al., (2017) who reported increased lymph node invasion and metastatsis in 30% of sunitinib-treated renal cell carcinoma patients via VEGFC-dependent lymphangiogenesis. Is there a process to share references? Colwiz? Aedin

Topic suggestion: Heterogeneity of metastatic spread page is loading…