hammerlab / immune-infiltrate-explorations

5 stars 1 forks source link

Generate data at MUSC #14

Open hammer opened 7 years ago

hammer commented 7 years ago

Some questions

hammer commented 7 years ago

Bob Wilson mentioned we should sort directly into Trizol; also mentioned at Does somebody have expertise in RNA isolation after FACS sorting? on ResearchGATE and in UW's Cell sorting for RNA isolation

hammer commented 7 years ago

Unfortunately there's no mention of how the cells were handled post-sorting and pre-RNA extraction in De novo transcriptome profiling of highly purified human lymphocytes primary cells.

They do list their markers though in Table 1: screen shot 2016-12-06 at 4 38 12 pm

hammer commented 7 years ago

Effects of RBC removal and TRIzol of peripheral blood samples on RNA stability (2011) has some evidence that Trizol is useful.

Some papers that cite it and also may be useful:

hammer commented 7 years ago

Came across this nice protocol that uses FISH to sort on RNA rather than cell surface markers: Transcriptional profiling of cells sorted by RNA abundance (2014)

hammer commented 7 years ago

CSHL Trizol protocol: Purification of RNA Using TRIzol (2010)

hammer commented 7 years ago

Use of Trizol with small amount of sample material: A simple and loss-free method to remove TRIzol contaminations from minute RNA samples (2009)

hammer commented 7 years ago

From Michelle Nelson in Chrystal's lab

I don’t sort into Trizol. The flow cytometer will sort cells using FACS buffer and will therefore add volume. Instead I sort into culture media and then just following I spin the cells down and re-suspend in Trizol. If you want to send me the link to the CSHL protocol I can look it over. I’m not sure if the RNA of interest will be altered by letting them sit longer, but the act of staining, washing, sorting will probably change them a bit already.

hammer commented 7 years ago

From Chrystal:

Michelle sorts her various human CD4+ T cell subsets and then adds Trizol afterwords. She doesn't add Trizol while the cells are being collected into tubes. See her below email.

We can not open the link to the CSHL protocol - it says that the link has been disabled to protect security. If you have the protocol handy or another link to it, we would be happy to take a look at it. We are always happy to find better ways to do an experiment.

Michelle says that the flow sorter (FACS Aria) is busy next week but mostly empty the following week. Make sure to sign up for sort time in advance. We can sign up for you to hold the spot if you want. We can also order normal donor buffys.

Finally, do you want to sort human CD8 T cells into naive, stem, central, effector and terminal memory compartments? If so, do you know what markers you want to use to sort them? We may have most of the flow antibodies for you to use, but If not we can quickly purchase them.

hammer commented 7 years ago

From Chrystal:

Do you know the frequency of DC subsets you expect to find in the peripheral blood of normal donors? Our expertise is more with T cell subsets and we know how many we can usually obtain from blood. We have done some work with DCs in mouse models or the use of artificial APCs to activate human T cells. Regardless, the process of sorting DCs should involve the same basic protocols.

It might be wise to first determine the frequency of DCs (based on your markers) in the blood before sorting them. Also, many labs will program immune cells into DCs with GMCSF/IL4 and then sort. This process might generate more DCs for your purpose but they will be manipulated. How you want to do it boils down to what question you want to ask. Do you want to understand the biology of the naturally arising DCs in the blood or rather do you want to study manufactured DCs? We can chat about this on Wednesday. In the meantime, I will search for papers that enrich DCs from peripheral blood.

Michelle has ordered 2 buffys. She can help you with all of the questions you have regard how to become independent in this process. We will set you up. However, we are always happy to hel

hammer commented 7 years ago

Just doing some reading on what's in the buffy coat to see if we'll have DCs to sort, and trying to figure out what to sort on.

hammer commented 7 years ago

FluoroFinder can help us choose what to sort on

hammer commented 7 years ago

Maybe repeat this study w/ RNA-Seq? Phenotype, function, and gene expression profiles of programmed death-1(hi) CD8 T cells in healthy human adults (2011). Unfortunately doesn't seem too interesting as the PD-1hi CD8+ T cells they isolate from healthy subjects aren't functionally exhausted and appear to be effector memory cells.

hammer commented 7 years ago

DC markers

hammer commented 7 years ago

From the IRIS paper and their 2009 SLE application.

screen shot 2016-12-06 at 9 20 51 pm

screen shot 2016-12-06 at 9 42 57 pm

T-cell subset isolation

CD8+

CD4+

Memory

Dendritic cell subset isolation

hammer commented 7 years ago

From the CIBERSORT paper, where 22 cell types are identified. The spreadsheets (supplementary items 2 and 3) have the useful information.

screen shot 2016-12-06 at 9 31 22 pm

screen shot 2016-12-06 at 9 48 32 pm

Non-IRIS data sources (LM22)

Validation data sources

hammer commented 7 years ago

Consider MDSCs? This paper claims that G-MDSCs correlate best w/ tumor burden.

Toward harmonized phenotyping of human myeloid-derived suppressor cells by flow cytometry: results from an interim study (2016) makes it clear this will not be easy.

262_2015_1782_fig1_html

hammer commented 7 years ago

Sorting strategy from CD39 Expression Identifies Terminally Exhausted CD8+ T Cells (2015):

journal ppat 1005177 s003

hammer commented 7 years ago

Just in case we decide to do some ultra low input sequencing: Evaluation of commercially available RNA amplification kits for RNA sequencing using very low input amounts of total RNA (2015)

hammer commented 7 years ago

Also I should note here that Miriam Merad from Sinai said she could send us their DC panel and is going to connect me to Christophe Benoist who does ultra-low input RNA-Seq for the ImmGen project.

hammer commented 7 years ago

Just noticed that the ImmGen project publishes their protocols:

They also have an 11-cell RNA-Seq mouse project that may be useful for us.

hammer commented 7 years ago

From Chrystal:

We use buffy coats from our Penn connection as they are much cheaper ($50) than Leuko Paks ($1000). We have a protocol for harvesting the PBMCs from the buffy we can share with you. For our purposes, we get enough T cells for our in vitro experiments. If we need more T cells for in vivo mouse work, we sometimes get a Leuko Pak. Michelle knows the difference between the yields of T cells between these sources and we can talk about that during our meeting. This info might be insightful for extrapolating DC yield.

Getting enough yield of your desired DC subsets for RNA-Seq might be challenging. I think there are ways to amplify your RNA though. Perhaps conducting a pilot study with Michelle to execute the protocols would be useful. This way, with those two buffys you, can QC the frequency of your DC subsets by the flow in our lab (not sure which DCs you want to sort?). If the numbers look good we can go for it. Otherwise, until we get the logistics worked out for the DC project, we could sort CD8 T subsets for RNA-Seq. I think some of your team was interested in this project.

If the DC yield is too low with our buffys, you can order a Leuko Pak from Research blood components. You can request HLA-A2 Pak for $1222.

Link: http://researchbloodcomponents.com/products.html

Bei Liu (across from me) has sorted mouse DCs at MUSC. She might be helpful for this project. I don't think she has worked with human DCs though.

hammer commented 7 years ago

From Michelle:

I just want to make it clear, the flow core does not provide the flow antibodies. We will supply those for now and be executing all the staining. Therefore we can target CD39 or others if you like as long as we design a workable panel. The fluorofinder is an excellent resource for making new panels and there are links to get the catalog #’s for the antibodies you choose. I’ll double check with our research specialist (Megan) who does our ordering, but we get 30-50% off list price on the antibodies plus free shipping. The core only supplies reagents to run the sorter.

hammer commented 7 years ago

Chrystal sent me LC Sciences, a vendor in Houston that does scRNA-Seq.

hammer commented 7 years ago

Chrystal pointed me to The Chemokine Receptor CX3CR1 Defines Three Antigen-Experienced CD8 T Cell Subsets with Distinct Roles in Immune Surveillance and Homeostasis, a recent (published yesterday!) paper from Ulrich von Adrian that defines a new subpopulation of CD8+ T cells, peripheral memory T cells. We could consider sorting and profiling these cells w/ RNA-Seq.

jburos commented 7 years ago

To address the question of where Cibersort does particularly poorly, @maximz's work evaluating Cibersort is (I think) summarized in this notebook.

For example, this table summarizes error from Cibersort for different cell types :

screen shot 2016-12-07 at 4 41 46 pm

This would suggest the following cell types of particular interest:

  1. Naive B cells
  2. T regs
  3. T cells follicular helper
  4. B cells memory
  5. T cells CD4 memory resting
  6. NK cells activated
  7. T cells CD8

Most of these (with possible exception of Naive B cells) are pretty important for cancer.

hammer commented 7 years ago

Thought from @jburos: can we predict PD-1hi vs. PD-1lo for CD8+ T cells from their RNA-Seq profile?

jburos commented 7 years ago

This may or may not have some overlap with the exhausted T cell phenotype mentioned above

hammer commented 7 years ago

Just thought of another cell type to add: tumor-educated platelets (TEPs)

jburos commented 7 years ago

The other cell type mentioned in the Orchestration and Prognostic Significance of Immune Checkpoints in the Microenvironment of Primary and Metastatic Renal Cell Cancer as being prognostic in ccRCC are LAG-3+ CD8+ T cells

ChrystalPaulos commented 7 years ago

You could compare PD1+cd8+ T cells to LAG3+CD8+ T cells to cD39+CD8+ T cells. Also, does there exist of CD8 T cells in the blood that express all there of the markers? If so, are those super exhausted? Control is bulk CD8 T cells sent through the machine but not bifurcated on.

maximz commented 7 years ago

To add to @jburos 's post:

The common cell type confusions I've seen in cibersort were:

Here's a visual description of similarity in the Cibersort dataset (LM22):

image

hammer commented 7 years ago

@ChrystalPaulos yes that would be very interesting. It would be very nice to be able to tease apart CD8+ exhausted, anergic, and senescent T cells. Do you happen to know of a panel for those kinds of cells, or know of their frequency in healthy human subjects?

michellehnelson commented 7 years ago

Here is our basic protocol for isolating WBC from buffy/leukopaks that we will be following next week.

Lymphocyte Isolation from Buffy Coat.docx

ChrystalPaulos commented 7 years ago

@hammer, I will dig through the literature for frequencies of these markers on hCD8 T cells. at quick glance, I notice that PD-1 is on peripheral CD8s at around 2-10 %. From our own experience, the expression of these markers can be vary from donor to donor. Another thought is I wonder if the "expected" biology of PD-1+cd8 T cells in the blood are not being unmasked by RNA-seq because they are not being stimulated via their TCR? We could simulate this generically by activating with a CD3 agonist post sort. That would mean not adding Trivol immediately. Lots of variables...

hammer commented 7 years ago

@ChrystalPaulos @michellehnelson the HIMC at Sinai has suggested the Ovation® Ultralow Library System V2. Have y'all heard of it/had any experience with it?

They also pointed me to Protocols for the Identification and Isolation of Antigen-Presenting Cells in Human and Mouse Tissues for DC sorting ideas. Unfortunately this chapter covers DCs from non-lymphoid tissue, not blood. Nevertheless they give a panel for skin (CD45, HLA-DR, CD14, CD1a, CD1c, CD11c, CD141) and soft tissue (CD3, CD19, CD20, CD56, CD123, HLA- DR, CD14, CD1c, CD11c, CD141).

screen shot 2016-12-08 at 3 44 38 pm screen shot 2016-12-08 at 3 44 51 pm

hammer commented 7 years ago

Bob Wilson from our MUSC core said he's worked with Clontech SMART-Seq® v4 Ultra® Low Input RNA Kit. Library prep is $320, then sequencing is $515 for 100MM 125 bp PE reads.

hammer commented 7 years ago

Also I found this Blood Dendritic Cell Enumeration Kit, human; @michellehnelson any chance our core has access to this kit?

The Anti-BDCA Cocktail is a pre-mixed cocktail which includes all antibodies needed for direct identification of PDCs, MDC1s, and MDC2s by flow cytometry. The DC-specific antibodies CD303 (BDCA- 2), CD1c (BDCA-1), and CD141 (BDCA-3) are conjugated to different fluorochromes and are detected in separate fluorescence channels of the flow cytometer. Apart from the DC-specific antibodies, the Anti-BDCA Cocktail contains monoclonal antibodies directed against CD19 for exclusion of B cells and CD14 for exclusion of monocytes. B cells and monocytes are excluded from the analysis as a subpopulation of B cells expresses CD1c (BDCA-1), and monocytes express CD141 (BDCA-3) at a low level. Both antibodies are conjugated to PE-Cy5.

hammer commented 7 years ago

From Adeeb at the Sinai HIMC:

It’s usually most efficient to do a 2 step process of initial MACS enrichment followed by specific DC subsets sorting by FACS. A minimal panel should include:

Dump channel with CD19, CD3 and CD56 (CD19 is most important) CD14 CD16 HLADR CD1c CD141 CD123 CD11c (not critical)

ChrystalPaulos commented 7 years ago

@hammer, you will have to buy this Blood Dendritic Cell Enumeration Kit from MACs. We do not have a core at MUSC that stocks any of these reagents. Michelle will show you how she uses such a kit to enumerate human T cells - we get those kits from MACs or R&D. Deresa Teal and or James Fant can help you get started on Market Place to order these reagents. Megan can also describe the process of getting such reagents for your lab.

ChrystalPaulos commented 7 years ago

We use Clontech SMART-Seq® v4 Ultra® Low Input RNA Kit. Have not compared it to the Ovation kit.

hammer commented 7 years ago

Review paper recommended by Danny Khalil at MSK with DC panel suggestions: Cross-Presentation in Mouse and Human Dendritic Cells (2015)

hammer commented 7 years ago

The final decision for next week thanks to @ChrystalPaulos and @michellehnelson: FITC- PD1, PE-CD3, PE Cy7-Lag3, APC-CD39, V450- CD4 and V500-Live/Dead.

ChrystalPaulos commented 7 years ago

Thanks! We can look into it.


From: Jeff Hammerbacher [notifications@github.com] Sent: Friday, December 09, 2016 4:35 PM To: hammerlab/immune-infiltrate-explorations Cc: Paulos, Chrystal; Mention Subject: Re: [hammerlab/immune-infiltrate-explorations] Generate data at MUSC (#14)

Review paper recommended by Danny Khalil at MSK with DC panel suggestions: Cross-Presentation in Mouse and Human Dendritic Cellshttps://www.ncbi.nlm.nih.gov/pubmed/26073982 (2015)

� You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/hammerlab/immune-infiltrate-explorations/issues/14#issuecomment-266129884, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AXT7dpvvBfE0Ok8T2flSO4c_GgahcJ81ks5rGcmigaJpZM4LAg1e.


This message was secured via TLS by MUSC.

hammer commented 7 years ago

@ChrystalPaulos pointed me to some of her previous work: A human memory T cell subset with stem cell-like properties (2011). They identify T stem cell memory (T_SCM) cells as a subset of memory T cells and do microarray expression profiling on them.

ChrystalPaulos commented 7 years ago

img_0077

It occurs to me (based on this 2011 R. Ahmed paper-snapshot above) that you should consider collecting the negative fraction of your sort for rna-seq! In other words, also take pd-1 negative cd8, lag-3 negative cd8, cd39 negative cd8 and if you can triple negatives cd8. This will be an additional 4 samples to you 5 samples all positive and bulks)- with 2 healthy donor Buffy's this would be 18 total. I think if all goes well, you can build on your data set with samples from cancer patients and hiv patients. Could be important contribution to field.

hammer commented 7 years ago

@ChrystalPaulos sounds great! We are definitely inspired by that paper (sent to me by Sasi in John Wherry's lab and mentioned above: https://github.com/hammerlab/immune-infiltrate-explorations/issues/14#issuecomment-265298414). Do you know anyone at MUSC who might have an HIV cohort?

ChrystalPaulos commented 7 years ago

Wei Jiang is a faculty in our department that might be helpful in advising you to an HIV cohort. Maybe Eric Meissner too.

BTW, we have been getting blood from SLE patients. Could be interesting.

finally, check out this site. We have not used them but allcell seems to have a collection of blood from patients with a variety of cancer disorders.

https://www.allcells.com/products/by-disease

hammer commented 7 years ago

@ChrystalPaulos when you say "samples from cancer patients", are you thinking of whole blood or TILs? I don't know much about the number of T cells pulled out of tumors but it would be really fascinating to look at their transcriptional profiles.

ChrystalPaulos commented 7 years ago

Sorry for lack of clarity. Looks like you can buy peripheral blood or bone marrow(?) from healthy as well as diseased individual at various volumes (40 to 400mls). You can pick they type of disease as well. They have services that supply plasma, B cell etc. They can derive blood from patients with various cancers as well as diabetes and chrones disease. You can see all of this one their website. Price is high but could be used for immediate work while you are setting up resources at MUSC. Also, Juan is working on getting peripheral blood and other tissues from his patients with heme cancers for our lab. We could likely share this with you once all paper work is approved. This could be good for this project too.

hammer commented 7 years ago

@michellehnelson what's the value of CD4 in our current panel?