Patent Application
20240093289
The present application is being filed with a Sequence Listing in electronic format. The Sequence Listing is provided in XML format as a file entitled “2023-10-7_125141.04359_WIPO_Sequence_Listing_XML.xml”, which is 4,492 bytes in size and was created on Oct. 7, 2023. The sequence listing is electronically submitted with this application via Patent Center and is incorporated herein by reference in its entirety.
Spatial transcriptomics is a rapidly advancing field, encompassing a range of different technologies and capable of spatially-resolved gene expression analysis. In contrast to single-cell RNA sequencing, which provides deep insights into heterogeneities within cell populations but does not preserve the spatial relationship among individual cells, techniques such as MERFISH, seqFISH, and Slide-seq can link these heterogeneities to differences in spatial composition and cellular proximity. However, apart from Niche-seq, they have either been applied only in vitro, or rely on generating tissue sections, which has confined their applicability to tissues that are easily sectioned. They map the gene expression profiles onto two-dimensional images, and extrapolation to three-dimensional architecture of intact tissues is still limited. For example, 3D expression profiles and cell segmentation have been demonstrated in tissue sections but the section thickness is limited by mRNA probe diffusion. In addition, tissue sections can provide only static images, necessitating the use of indirect methods such as pseudo-time analysis to infer cellular trajectories over time, and none of the currently available spatial transcriptomics and multi-omics technologies have been combined with in vivo imaging.
The present disclosure provides systems and methods that overcome one or more of the aforementioned drawbacks via methods and systems that enable image-guided cell isolation for scRNA-seq and multi-omics analysis, which may be referred to as Image-seq herein. The core of the Image-seq system may include a one-photon or multiphoton microscope with two optical paths, one for imaging and one for laser micromachining, or a single optical path for both imaging and micromachining which creates an access channel in tissue through which a micropipette is brought to the target location and aspirates cells under image guidance. Because it captures viable cells, Image-seq can be combined with state-of-the-art library preparation protocols, leading to higher mRNA detection efficiencies and broader transcript coverage than other spatial sequencing technologies. In addition, standard computational tools can be used for data analysis.
According to one aspect of the present disclosure, a method for image-guided cell isolation from a region of interest of a subject is disclosed. The method includes imaging the subject using optical microscopy to identify the region of interest in a target anatomy, inserting a micropipette into the region of interest under guidance of the optical microscopy, and aspirating at least one cell of a target population of cells in the region of interest under guidance of the optical microscopy. The method further includes analyzing the at least one cell aspirated from the region of interest.
According to another aspect of the present disclosure, a method for image-guided tissue ablation from a region of interest of a subject is provided. The method includes the steps of imaging the subject using optical microscopy to identify a tissue in a region of interest of a target anatomy and ablating the tissue in the region of interest using an ablation laser under guidance of the optical microscopy. The method further includes confirming a spatial location of the ablated tissue using the optical microscopy.
According to another aspect of the present disclosure, a system for image-guided cell isolation from a region of interest of a subject is provided. The system includes an optical microscope, a flushing assembly, a micropipette assembly, and a processor in communication with the optical microscope and the micropipette. The processor is configured to image the subject using the optical microscope to identify the region of interest in a target anatomy, identify from one or more images generated of the subject at least one cell of the target population of cells in the region of interest under guidance of the optical microscope, and control the micropipette assembly to aspirate at least one cell of a target population of cells in the region of interest under guidance of the optical microscope.
These aspects are nonlimiting. Other aspects and features of the systems and methods described herein will be provided below.
The foregoing features of embodiments will be more readily understood by reference to the following detailed description, taken with reference to the accompanying drawings, in which:
As used in this specification and the claims, the singular forms “a,” “an,” and “the” include plural forms unless the context clearly dictates otherwise.
As used herein, “about”, “approximately,” “substantially,” and “significantly” will be understood by persons of ordinary skill in the art and will vary to some extent on the context in which they are used. If there are uses of the term which are not clear to persons of ordinary skill in the art given the context in which it is used, “about” and “approximately” will mean up to plus or minus 10% of the particular term and “substantially” and “significantly” will mean more than plus or minus 10% of the particular term.
As used herein, the terms “include” and “including” have the same meaning as the terms “comprise” and “comprising.” The terms “comprise” and “comprising” should be interpreted as being “open” transitional terms that permit the inclusion of additional components further to those components recited in the claims. The terms “consist” and “consisting of” should be interpreted as being “closed” transitional terms that do not permit the inclusion of additional components other than the components recited in the claims. The term “consisting essentially of” should be interpreted to be partially closed and allowing the inclusion only of additional components that do not fundamentally alter the nature of the claimed subject matter.
The phrase “such as” should be interpreted as “for example, including.” Moreover, the use of any and all exemplary language, including but not limited to “such as”, is intended merely to better illuminate the invention and does not pose a limitation on the scope of the invention unless otherwise claimed.
Furthermore, in those instances where a convention analogous to “at least one of A, B and C, etc.” is used, in general such a construction is intended in the sense of one having ordinary skill in the art would understand the convention (e.g., “a system having at least one of A, B and C” would include but not be limited to systems that have A alone, B alone, C alone, A and B together, A and C together, B and C together, and/or A, B, and C together.). It will be further understood by those within the art that virtually any disjunctive word and/or phrase presenting two or more alternative terms, whether in the description or figures, should be understood to contemplate the possibilities of including one of the terms, either of the terms, or both terms. For example, the phrase “A or B” will be understood to include the possibilities of “A” or “B” or “A and B.”
All language such as “up to,” “at least,” “greater than,” “less than,” and the like, include the number recited and refer to ranges which can subsequently be broken down into ranges and subranges. A range includes each individual member. Thus, for example, a group having 1-3 members refers to groups having 1, 2, or 3 members. Similarly, a group having 6 members refers to groups having 1, 2, 3, 4, or 6 members, and so forth.
The modal verb “may” refers to the preferred use or selection of one or more options or choices among the several described embodiments or features contained within the same. Where no options or choices are disclosed regarding a particular embodiment or feature contained in the same, the modal verb “may” refers to an affirmative act regarding how to make or use an aspect of a described embodiment or feature contained in the same, or a definitive decision to use a specific skill regarding a described embodiment or feature contained in the same. In this latter context, the modal verb “may” has the same meaning and connotation as the auxiliary verb “can.”
System Component
In one aspect, the disclosed system comprises an optical and mechanical apparatus that can perform both multiphoton and/or confocal imaging (at arbitrary frame rates) as well as plasma-mediated laser ablation. This makes it possible to aspirate/isolate cells with a micropipette under image guidance from small, spatially confined volumes of mouse bone marrow tissue (on the order of 50×50×50 μm, although the exact volume can be controlled by the device and can be significantly smaller or larger). Note that in tissues like the bone marrow, plasma-mediated laser ablation is required to perform a microsurgery procedure on bone and generate a small opening or access port to the bone marrow underneath. This enables insertion of the micropipette so that cells can be aspirated under image guidance. In other tissues the laser ablation procedure could be used to dissect out small tissue blocks. Further tissue types may not require laser ablation at all and cells could be aspirated directly by micropipette. In all the above settings, the three-dimensional (3D) spatial architecture is first visualized by multiphoton and/or confocal microscopy, the device then isolates cells (that could be viable or fixed) from a precise anatomical location. By imaging the tissue both before and after cell aspiration, the precise 3D spatial position of the isolated cell sample is determined. Through the analysis of these spatially-harvested micropipette samples, it is possible to study spatial variations in cell type, cellular function and cell-cell communication. In a disease setting such studies are important for the development of new therapeutic targets. One of the analysis tools (single-cell RNA-sequencing) is described in greater detail in the Examples below. It is this combination of cell harvest by micropipette under image guidance with single-cell RNA-sequencing that defines a new technology for spatially-resolved transcriptomics.
As illustrated in
Further, the processor 112 may identify the region of interest from the images generated from the subject. In a non-limiting example, the region of interest includes at least on cell of a target population of cells. As will be described in the Examples below, the target population of cells may include bone marrow stromal cells or acute myeloid leukemia (AML) cells in the bone marrow cavity of a bone. In a non-limiting example, the target population of cells include a label that is identifiable using the optical microscope 102. For example, the label may include a fluorescently tagged molecule that is expressed by or selectively bound to the target population of cells. Examples of labels are provided in the Example section below. In a non-limiting example, the micropipette tip is also coated with a label for visualization under guidance of the optical microscope.
In a non-limiting embodiment, the optical microscope may be further integrated with an ablation laser 104. The ablation laser 104 can be the same as the multiphoton imaging laser. However, it should have enough power to generate a plasma in the microscope's image plane (which requires a pulse energy of at least 10-20 nJ). Optimal pulse energies will vary depending on the exact properties of the optical system (including laser repetition rate and pulse length), as well as the tissue type and exact application.
In a non-limiting example, the processor 112 is further configured to control the micropipette assembly for inserting the micropipette 106 into the region of interest.
As shown in
In a non-limiting example, the ablation laser emits the ablation beam at a repetition frequency range of 1 kHz to 10 MHz.
In a non-limiting example, the ablation laser emits the ablation beam at a pulse energy in a range of about 5 nJ to 1 μJ.
In a non-limiting example, the ablation laser emits the ablation beam at a pulse duration of less than 1 ps.
As will be described in further detail below, the ablation laser 114 may be used to generate an opening in the target anatomy. For example, to access the target population in a bone marrow cavity, it is necessary to form a channel in the bone configured for insertion of the micropipette. The ablation laser 114 may alternatively or additionally be used to ablate the tissue in the region of interest.
The system 100 in
In another non-limiting example, the system 100 in
In another non-limiting example, the system 100 in
In a non-limiting example, the system 100 is modular and may include any number and combination of the elements described above.
Method
Single-cell RNA sequencing of spatially-harvested cell samples using the Image-seq device described above represents a new technology for spatial transcriptomics. Spatial transcriptomics includes several commercially available technologies (for example, 10× genomics' Visium or Vizgen's Merscope platform), as well as a number of technologies that are still in development and not yet commercially available (for example Slide-Seq). Because of Image-seq's ability to isolate intact (viable or fixed) cells it can be combined with state-of-the-art library preparation kits and protocols that are commercially available. When combined with Smartseq library preparation it has a higher sensitivity (number of genes that are detected per cell) than any currently available spatial transcriptomics technology. Image-seq can also be used to study rare cells that would not be detected by other spatial transcriptomics technologies (see Example 1 below) and does not require the generation of tissue sections, which is a limitation of commercially available technologies. Image-seq may be combined with state-of-the-art multi-omics analysis tools like (single-cell) proteomics, metabolomics, ATAC-seq and DNA sequencing. Combinations of the device with these single-cell analysis tools represent a new technology for spatially-resolved single-cell multi-omics.
The advantages of Image-seq include the combination of multiphoton/confocal microscopy with spatially-resolved single-cell RNA sequencing, the method's high sensitivity, its applicability to rare cells and stromal cells, as well as its potential for multi-omics analysis. In addition, it does not require the generation of tissue sections, making it the only spatially-resolved single-cell RNA sequencing technology that can currently be applied to bone marrow tissue.
As described in the Example section below, the systems and methods disclosed herein generate high-quality single-cell RNA sequencing data. It has been shown that it can be used to study hematopoietic cells, stromal cells, rare cancer cells, and that it provides higher sensitivity than any currently available spatial transcriptomics technology. It has also been demonstrated that it doesn't require the generation of tissue sections and can be used in conjunction with other single-cell analysis techniques (i.e. flow cytometry and microfluidic cell sorting).
In another aspect of the present disclosure, a method of identifying and aspirating target cells is disclosed. The method illustrated in
The method 200 of
In a non-limiting example analyzing the aspirated cells at step 208 includes, but is not limited to, spatially resolved transcriptomic analysis, single cell RNA sequencing, bulk RNA sequencing, single-cell multi-omics analysis, bulk multi-omics analysis, genetic manipulation of aspirated the cell(s), transplantation of the aspirated cell(s), and cell colony formation. The Example sections below provide disclosure for single cell RNA-sequencing and spatially resolved transcriptomic analysis.
In a non-limiting example, multi-omic analysis may be performed using droplet based single cell multi-omics sequencing or multi-omic analysis with flow cytometry sorting.
In a non-limiting example, transplantation may involve transplanting spatially harvested sample into irradiated or non-irradiated recipients. The cells of the samples may or may not be (i) sorted prior to transplantation using either a flow cytometer or microfluidic cell sorter and (ii) genetically manipulated (for example gene knockout or gene knockdown) prior to transplantation. Such studies may aid in revealing functional differences between cells from different spatial locations.
In a non-limiting example, colony formation assays involve in vitro culturing of spatially-harvested samples (as described in the Example section below).
In another aspect of the present disclosure, a method of ablating tissue is disclosed. The method illustrated in
The method 210 of
In a non-limiting example, method 210 of
In a non-limiting example, methods 200 and 210 of
It will be appreciated by those skilled in the art that while the disclosed subject matter is described above and in the examples below in connection with particular embodiments and examples, the invention is not necessarily so limited, and that numerous other embodiments, examples, uses, modifications and departures from the embodiments, examples and uses are intended to be encompassed by the claims attached hereto. Each article cited herein is incorporated by reference in its entirety.
Various features and advantages of the invention are set forth in the following examples.
The Image-seq methods and system described above have been used to study the bone marrow, the primary site of hematopoiesis, where hematopoietic stem cells give rise to all of the blood cells in the body. The bone marrow is also the site where malignant cells can either originate (such as leukemia) or preferentially metastasize to (such as prostate or breast cancer). Although the functional organization of hematopoietic cells in the bone marrow has been extensively characterized, their 3D spatial organization has been difficult to assess due to their complexity and location deep inside the bone matrix. To demonstrate Image-seq's versatility and high sensitivity, high-throughput, droplet-based sequencing was used with the 10× Chromium platform to profile BM hematopoietic cells, as well as the Smartseq v4 protocol (Takara Bio USA, Inc., Mountain View, USA) to profile rare (<0.01% leukemic burden) acute myeloid leukemia (AML) cells and bone marrow stromal cells. Specifically, AML progression was tracked using intravital microscopy, and observed pronounced spatial heterogeneity in the earliest stage of expansion from single AML cells seeded in the bone marrow. This highlighted the need to capture these cells from distinct bone marrow locations where they either proliferate or remain quiescent, in order to identify the factors that regulate leukemia progression in the earliest stage of the disease development. Using this approach, DPP4 was identified as a key upregulated gene in AML cells from the more proliferative bone marrow compartments. Strikingly, DPP4 was not expressed on the same cells cultured in vitro, suggesting that DPP4 was specifically activated in vivo and was correlated with disease progression.
Results
Cell Isolation Under Image Guidance
The core of the system is a confocal or multiphoton microscope with an additional laser beam capable of tissue ablation. The requirement for the ablation laser is that it needs to have sufficient pulse energy (>10 nJ per pulse, or approximately 10-fold the pulse energy typically used in multiphoton microscopy) to generate a plasma through multiphoton ionization at the laser focus. To minimize thermal damage, it also needs to have a low average power, which can be accomplished by reducing the pulse repetition frequency from the ˜80 MHz typically used for multiphoton microscopy to a few MHz to compensate for the higher pulse energy needed for tissue ablation. These requirements may be readily fulfilled with commercial or industrial femtosecond fiber lasers for micromachining or ophthalmic microsurgery. The plasma is used to ablate or etch away bone with minimal collateral damage to the surrounding tissue. A small, ˜50×100 μm channel is created, through which a micropipette, controlled by a micropipette assembly or micromanipulator, is inserted to aspirate live bone marrow cells under image guidance. The sample is expelled from the micropipette, which directly generates a single-cell suspension.
In the implementations presented herein, the imaging arm and the ablation arm are powered by a single femtosecond fiber laser operating at 5 MHz repetition frequency. This repetition frequency enables full-field (500×500 pixels) image acquisition at 15 frames per second (half the video rate) with a single laser pulse per pixel. However, it is also possible to integrate an ablation capability into an existing confocal or multiphoton microscope by adding an ablation laser that fulfills the requirements described above. Detailed optical designs are shown in
To further demonstrate the system's ability to isolate live cells from defined anatomic regions of murine bone marrow, it was used to isolate bone marrow samples from β-actin-green fluorescent protein (β-actin-GFP) mice.
By perfusing with enzymatic digestions buffer after the initial perfusion (to reduce the number of red blood cells) and incubating the bone sample for −20 min, a protocol for the isolation of stromal cells was established.
Validation of Image-Seq Technology
An outline of the Image-seq workflow is shown in
A single-cell suspension is generated by expelling the sample from the micropipette into a tube that is transferred directly to the 10× Chromium chip for high-throughput characterization of the entire cell sample. Alternatively, samples can be stained, and individual cell populations sorted into wells by flow cytometry, for subsequent library preparation and sequencing using the Smartseq-v4 protocol. To validate the technology, 11 Image-seq samples from a total of n=5 C57B16 mice were collected, along with three whole calvarium bone marrow (WCBM) samples (n=3 mice) for single cell isolation and library preparation using the 10× Chromium platform. After sequencing and read alignment, the Conos package in R was utilized to integrate the multiple scRNA-seq datasets and align them with other public scRNA-seq data. Leiden clustering was used to determine joint cell clusters across the entire dataset, identifying most major hematopoietic cell populations, and visualizing them by uniform manifold approximation and projection (UMAP) embedding (
As expected for sampling small volumes from different locations, significant sample-to-sample variations were observed (a detailed breakdown is given in
To demonstrate the technology's ability to obtain spatially-resolved single-cell transcriptional data of stromal cells (
To further validate the spatial selectivity of the technology, an MLL-AF9 mouse model of acute myeloid leukemia (AML) was imaged in which leukemia cells express GFP under the ubiquitin promoter at day 10 after transplantation. As shown previously, regions of high leukemic burden were found to be interspersed with regions of low leukemic burden (
Image-Seq Analysis of Early Leukemia Expansion
Next, early leukemia progression in a Hoxa9/Meis1-Ubiquitin-c-GFP (HA9M1) mouse model of AML was evaluated by performing intravital imaging of the calvarium bone marrow between 1 and 3 days after transplanting 3×106 cells into non-irradiated recipients (
P, NP and IM AML cells were identified by imaging and aspirated by micropipette using the Image-seq platform. Subsequently, the AML cells were separated from the ˜100-400 surrounding hematopoietic cells by sorting them into individual wells of a 96-well plate by flow cytometry and gating for GFP (see
After sequencing and read alignment, cells that were in the G0 phase were identified and hierarchical clustering was performed based on cell cycle genes, identifying three separate clusters (shown in the heatmap of
Next, AML cells were sub-clustered after regressing out cell cycle genes using Seurat, with the resulting UMAP embedding shown in
Increased proliferation of DPP4-positive cells in mouse and human AML
With the goal of identifying signals that trigger the exit of leukemia-initiating cells from the quiescent state, the differential expression of genes between P and NP cells was investigated (
To validate DPP4 expression its protein expression was assessed using flow cytometry and found that it increases as a function of disease progression, with a maximum stable expression at ˜50% by week 4 (
Because DPP4 expression was confined mainly to the AML-mono cluster, which expressed high levels of Itgb7, Flt3 and Cd48 (
To assess leukemia-initiating capacity, 1,000 DPP4high and DPP4neg AML cells were isolated 3 weeks after transplantation using flow cytometry. They were transplanted into fresh, non-irradiated recipients and compared leukemic burden and the proportion of DPP4-positive and DPP4-negative cells at day 10 (
Activation of DPP4 Expression by the Microenvironment
While observing DPP4 cell surface expression to be highly stable and reproducible in vivo, no DPP4 on AML cells cultured in vitro were detected (
AML is an aggressive blood cancer characterized by an accumulation of immature myeloid cells in the bone marrow that are arrested in differentiation and that accumulate as immature and malignant self-renewing progenitors. Despite an initially favorable response to treatment with intensive cytotoxic chemotherapy, ˜75% of patients die within 5 years of diagnosis. Relapse is thought to be driven by the rare pool of leukemia-initiating cells that persist in the bone marrow following chemotherapy. An important approach to the development of new therapeutics is the targeting and exploiting of the supportive interaction as leukemia-initiating cells communicate with and seek refuge within the bone marrow microenvironment. Studying and identifying signals that trigger the exit of leukemia-initiating cells from the quiescent state thereby provides mechanistic insight into disease recurrence. Examining the bone marrow at a stage of very low leukemic burden, comparable to the state of minimal residual disease (when the AML cells are more likely to be surrounded by normal hematopoietic and stromal cells), provides unique insights that cannot be obtained using traditional assays of relatively high leukemic burdens in which the bone marrow is crowded with malignant cells and the microenvironment is severely altered.
The above results show how the combination of intravital microscopy to study the dynamics of AML disease progression, and spatially resolved scRNA-seq, provides new insights into leukemia biology. Moreover, multiphoton microscopy uniquely informs the 3D spatial context of AML cells and can be used to validate the single-cell gene expression data in a spatially resolved manner (
In situ imaging of murine bone marrow has led to numerous insights into the basic biology of hematopoietic stem cells, as well as the spatial organization of bone marrow tissue. In the skull bone marrow, intravital microscopy has been used for the study of hematopoiesis, along with hematopoietic stem cell and leukemia biology. It has elucidated the temporal dynamics of these processes and tracked the association of individual cells with distinct bone marrow components in real time. Imaging alone, however, fails to provide unbiased mechanistic insight into the observed cellular dynamics and spatial organization. It is precisely such information, however, that promises to bring new insights into hematopoietic and leukemia biology and, concomitantly, the development of new therapeutics.
To date, the only spatially resolved transcriptional study of the murine bone marrow, a tissue that remains highly challenging to section, relied on bulk transcriptomic analysis of tissue blocks that were cut from formalin-fixed bone marrow sections using laser micro-dissection. The cellular composition of individual blocks was then inferred computationally using a separate, scRNA-seq dataset. At present there have been no spatially resolved transcriptional studies of the leukemic bone marrow.
Image-seq represents a new experimental approach for integrating spatial and molecular information. Multiple contrast mechanisms can be used to visualize the procedure and reconstruct the 3D spatial position of the extracted cell sample, including autofluorescence, confocal reflectance (
Microscope
Intravital microscopy and plasma-mediated laser ablation were performed using a custom-built multiphoton microscope. The output of a femtosecond, 1,550 nm fiber laser (Calmar Cazadero) operating at a repetition rate of 5 MHz was split into two optical paths: one was frequency doubled with a BiBO crystal (Newlight Photonics) to obtain a wavelength of 775 nm that was used for either imaging or ablation (
Procedure for Intravital Imaging
The procedure for intravital imaging of the calvarium bone marrow is described in detail elsewhere 61. Prior to intravital microscopy and ablation, analgesics were given (buprenorphine at 0.05-0.1 mg kg−1 i.p.), mice were anesthetized using vaporized isoflurane (3-4% for induction, 1-2% for maintenance) and depth of anesthesia ensured by toe pinch. Hair around the incision site on the scalp was trimmed, and skin was made aseptic using a betadine scrub. The incision (˜5 mm×7 mm) was made using sterile surgical scissors, and the skin folded back to expose the skull bone, which was hydrated using sterile PBS. Mice were transferred to a mouse holder with integrated heating pad (37° C.), and a continuous stream of isoflurane supplied via a nose cone during in vivo imaging and ablation. Intravital microscopy experiments were carried out using the microscope and dichroic mirror-filter configuration detailed above, and image stacks were acquired with a 2 μm step size from the calvaria surface and by averaging 15 frames to obtain a single image. At the end of each imaging session the mice were either sacrificed or survival surgery was performed. For survival, the exposed skull was extensively irrigated with sterile saline and the scalp closed with surgical sutures (Ethicon 6-0 nylon monofilament, Ethicon). After closure, 0.25% bupivacaine (2 mg per kg animal weight) was administered to the surgical site via percutaneous infiltration to aid with pain management, and triple antibiotic ointment (Curad) was applied on the sutured area. The animal was returned to its cage and monitored until awake. Buprenorphine (0.05-0.1 mg per kg animal weight) was given i.p. or s.c. along with topical antibiotic ointment every 8-12 h for up to 2 days after surgery.
In Vivo Cell Aspiration
The site for cell extraction was identified by intravital imaging of the calvarium bone marrow (procedure detailed above). A volume of bone ˜40 μm×200 μm×300 μm was removed using laser ablation (pulse energy 14 nJ). A circular channel was etched (diameter ˜100 μm, depth 20-30 μm, pulse energy 10 nJ) by placing an iris in the intermediate image plane, and a micropipette (MPB-FP-20, Origio) was inserted through the channel and into the bone marrow using a micromanipulator (Sutter Instruments). The target cell was aspirated by suction with an Air Syringe (Cooper Surgical) and the procedure was visualized using a combination of multiphoton and confocal reflectance signals. The ablation procedure itself was performed at a rate of 0.25 μm per 670 ms along the z dimension, which corresponded to 10 passes per plane using the 15 frame per second imaging rate of the optical system.
Image-Guided Cell Isolation for Flow Cytometry and Sequencing
In Vivo Imaging of Calvarium
Prior to cell isolation the mice underwent intravital imaging, the sites for cell extraction were identified and their spatial position recorded with respect to the bregma and lambda reference points. Mice were retro-orbitally injected with either Di-8-ANEPPS (1.9 mg kg−1) or Brilliant Violet 421 anti-mouse CD31 antibody (1 mg kg−1, BioLegend) to aid in the visualization of the cell isolation procedure. Anesthesia was increased to 4%, the mice were transferred to a dissection tray and were transcardially perfused as follows:
Mice were transferred back to the microscope and samples were sequentially isolated from positions marked for cell extraction.
In Situ Imaging of Tibia
Mice were perfused first with an ice-cold solution of 5 μM EDTA in PBS (ThermoFisher Scientific, flow rate 5 ml min−1, total volume 10 ml) and then with ice-cold PBS (ThermoFisher Scientific, no Ca or Mg, flow rate 5 ml min−1, total volume 10 ml). Tibia were then dissected and cleaned, and the tibial bone was thinned to a thickness of ˜50 μm using a razor blade. Bones were mounted onto a microscope slide by fastening a piece of modeling clay to the glass slide and gently pressing the bone onto the modeling clay. The mounted bone was then transferred to the microscope for in situ imaging.
Image-Guided Cell Aspiration (Calvarium and Tibia)
In each location the procedure was as follows: first, a volume of bone ˜40 μm×200 μm×300 μm was removed using a pulse energy of 14 nJ; second, a channel with dimensions ˜30 μm×50 μm×100 μm was created using a pulse energy of 10 nJ; third, the micropipette was inserted through this channel and positioned next to the target cells; and last, the cells were aspirated using a micropipette (MBB-FP-M-20, Origio) and transferred to an Eppendorf tube filled with 5 μl Medium-199 with 2% v/v FBS. Samples were kept on ice until they were either analyzed using flow cytometry (validation experiments), transferred to the 10× Chromium platform, or sorted into individual wells of a 96-well plate by flow cytometry (for library preparation by SMARTseq-v4). The ablation procedure itself was performed at a rate of 0.25 μm per 670 ms along the z dimension, which corresponded to 10 passes per plane using the 15 frame per second imaging rate of the optical system.
Prior to the experiment, micropipettes were coated with Sigmacote (flowed through the micropipette for 2 min at a rate of 200 μl min−1) to prevent cells from adhering to the glass surface, as well as with Qtracker 655 vascular labels (5 μl were pipetted up and down several times) to fluorescently coat the pipette and aid with visualization.
Collection of Whole Bone Marrow Preparations
Calvaria were dissected and cut into smaller pieces. Tibia were dissected and cleaned. To aid in the release of the bone marrow, calvaria bone fragments or whole tibia bones were gently crushed in Medium-199 (Gibco) supplemented with 2% FBS (Gibco). The resulting cell suspension was subsequently passed over a 70 μm cell strainer (BD Falcon).
Cell Lines
Syngeneic Leukemia Model
In brief, the HoxA9/Meis1 AML cell line was generated by retroviral transduction with an MSCV-HoxA9-IRES-Meis1 construct (originally designed by G. Sauvageau) into bone marrow mononuclear cells from a mouse expressing GFP under the control of the ubiquitin, and luciferase under the control of the β-actin promoter. The MLL-AF9 cell line used for the cell cycle experiments was generated by collecting bone marrow from a 5-fluorouracil-treated Cas9-GFP mouse, followed by two consecutive transfections with retroviral MLL-AF9. For both models the cells were transplanted into irradiated recipients, collected from terminally ill animals and re-transplanted into a second set of irradiated recipients, from which GFP-expressing cells were collected close to the disease endpoint. These cells were cultured in RPMI 1640 (Gibco) supplemented with 10% FBS (ThermoFisher Scientific), 100 IU ml−1 penicillin (Corning), 100 mg ml−1 streptomycin (Corning), 5 ng ml−1 interleukin 3 (IL-3, Peprotech), as well as 100 ng ml−1 stem cell factor (SCF, Peprotech) for the HoxA9/Meis1, and 20 ng ml−1 SCF and 10 ng ml−1 IL-6 (both from Peprotech) for the MLL-AF9. Recipient female mice (10-12 weeks old) were injected with 3×106 cells in 200 μl PBS (HoxA9/Meis1) and 1×106 cells in 200 μl PBS (MLL-AF9).
The MLL-AF9 model used for the experiments in
For transplantation of DPP4-negative and DPP4-positive HoxA9/Meis1 leukemia, cells were isolated from the long bones and vertebral column 3 weeks after transplantation. The marrow underwent density gradient centrifugation (Ficoll-Paque Plus, Cytiva Life Sciences) at 400×g for 25 min at room temperature with no brake. The mononuclear layer was isolated and subsequently blocked in PBS with 2% FBS and murine Fc Block (BD Biosciences, dilution 1:50). Following this, the samples were stained with CD45-APC/Cy7 (BD Biosciences, dilution 1:100) and DPP4-PE (Biolegend, dilution 1:20). To exclude dead cells, samples were incubated with 7-aminoactinmycin D (7AAD, 0.25 μg, BD Biosciences) and then the 7AAD−GFP+CD45+DPP4− and 7AAD−GFP+CD45+DPP4+ cells were sorted into separate tubes. A total of 1,000 cells of each phenotype were injected into 10-12-week-old recipient mice. Leukemia burden and DPP4 expression were assessed 10 days post-transplantation.
Simian Virus 40 Immortalized Bone Marrow Stroma
Total bone marrow cells were isolated from the femurs and tibias of B6J.129(B6N)-Gt(ROSA)26Sortm1(CAG-cas9*,−EGFP)Fezh/J mice (Jackson Laboratories). The bones were crushed in PBS (ThermoFisher) with 2% FBS (Gibco) and the released marrow was filtered over a 40 μm strainer. Mononuclear cells were collected by density gradient centrifugation (Ficoll-Paque Plus, Cytiva Life Sciences). A total of 20×106 mononuclear cells were put into culture with Alpha-MEM (Gibco) supplemented with 20% FBS (Gibco) and 1% Penicillin-Streptomycin (Gibco) in 150 mm dishes. Non-adherent cells were discarded around day 5 and the media changed every 5-7 days for approximately 3 weeks. At this point, colonies of large adherent fibroblasts were apparent. The cells were detached from the dishes with Trypsin-EDTA (Gibco), counted, and 50,000 cells seeded into two wells of a six-well plate. The following day, one well of cells was transduced with lentivirus in the presence of 8 μg ml−1 polybrene (Millipore, Sigma) to deliver the simian virus 40 (SV40) small and large T antigen. The plasmid pLenti CMV/TO SV40 small+Large T (w612-1) was a gift from Eric Campeau (Addgene plasmid 22298; http://n2t.net/addgene:22298; RRID: Addgene_22298). Despite the lack of a selectable marker, the transduced cells divided much more rapidly than the non-transduced primary stroma, and over the course of 2-3 passages, established an SV40 immortalized stromal cell line.
Isolation and In Vitro Stimulation of T Cells
Spleens were collected from C57B1/6 J mice and a single-cell suspension was obtained by mechanical dissociation of the tissue over a 70 μm cell strainer (BD Falcon) in RPMI 1640 (ThermoFisher) supplemented with 10% FBS (Gibco) and 1% Penicillin-Streptomycin (Gibco). Red cells were lysed using ACK Lysing Buffer (Quality Biological) followed by removal of non-T-cell splenocytes through magnetic depletion. In brief, the cell suspension was adjusted to 108 cells ml−1 and incubated with biotinylated antibodies directed against B220, CD19, Ter119, NK1.1, Cd11b and Gr1 (all from Biolegend, dilution of 1:100) at a concentration of 5 μg ml−1 for 10 min at room temperature on an orbital shaker. This was followed by the addition of 25 μl ml−1 streptavidin-conjugated Rapidspheres (Stem Cell Technologies) and an additional 5 min of incubation at room temperature on an orbital shaker. The samples were placed in an EasySep magnet (Stem Cell Technologies) for 5 min and the purified T cells were subsequently decanted into a fresh tube. Purity of the isolated T cells was confirmed to be >95% using FACS analysis. To induce T-cell proliferation, the isolated splenic T cells were plated with the T-cell activator Dynabeads CD3/CD28 (ThermoFisher) in a 1:1 ratio in leukemia cell line medium:RPMI 1640 (Gibco) supplemented with 10% FBS (Gibco), 100 IU ml−1 penicillin (Gibco), 100 mg ml−1 streptomycin (Gibco) and 5 ng ml−1 IL-3 (Peprotech), as well as 100 ng ml−1 SCF (Peprotech).
Flow Cytometry
Whole Bone Marrow and Micropipette Samples
Whole bone marrow and micropipette samples from calvarium and tibia were blocked with anti-mouse Fc Block (BD Biosciences, dilution 1:50) for 10 min at 4° C. The cells were thereafter stained with a blood cell lineage cocktail for 30 min at 4° C. For detection of dead cells 7AAD (BD Biosciences, 0.25 μg) was added to the sample prior to analysis. Flow cytometry was performed on a BD FACS Aria III sorter (BD Biosciences) and all data were analyzed using FlowJo (Treestar).
Flow Sorting of AML Cells for SMARTseq-v4
Prior to sorting, 1 ml PBS (ThermoFisher Scientific) was added to each sample tube, along with 0.1 μg DAPI (ThermoFisher Scientific). The sample was incubated for 10 min, gently vortexed and transferred to the flow cytometer (MoFlo Astrios EQ cell sorter). Single, live, GFP-positive AML cells (see
Flow Analysis and Sorting of CXCL12+ Stromal Cells for SMARTseq-v4
Prior to sorting or analysis, samples were cell-surface stained with anti-CD45-BV421 (BioLegend, dilution 1:100) for 30 min at 4° C. in Medium-199 (Gibco) supplemented with 2% FBS. Prior to sorting, 1 ml PBS (ThermoFisher Scientific) was added to each sample tube, along with 0.1 μg DAPI (ThermoFisher Scientific). The sample was incubated for 10 min, gently vortexed and transferred to the flow cytometer (MoFlo Astrios EQ cell sorter). Flow cytometry data were analyzed using FlowJo (Treestar). Single, live, CD45−DsRed+stromal cells (see description of gating strategy disclosed herein) were sorted into individual wells of a 96-well PCR plate filled with 2.6 μl Lysis Buffer (Takara Bio). Plates were sealed, spun down, snap-frozen and stored at −80° C. prior to preparation for cDNA synthesis using the SMARTseq-v4 assay.
Leukemia Burden and DPP4
Leukemic bone marrow was blocked using anti-mouse Fc Block (BD Biosciences, dilution 1:50) for 10 min at 4° C. in Medium-199 supplemented with 2% FBS. Surface staining was thereafter performed with CD45-APC/Cy7 (BD Biosciences, dilution 1:100) and DPP4-PE (BioLegend, dilution 1:20) for 30 min at 4° C. The cells were then washed and resuspended in Medium-199 supplemented with 2% FBS with 0.25 μg 7AAD (BD Biosciences). Flow cytometry was performed on a BD FACS Aria III sorter (BD Biosciences) and all data were analyzed using FlowJo (Treestar).
Leukemia Cluster Cell Surface Markers
Leukemic bone marrow was blocked using anti-mouse Fc Block (BD Biosciences, dilution 1:50) for 10 min at 4° C. in Medium-199 supplemented with 2% FBS. Surface staining was thereafter performed with a leukemia cluster cocktail for 30 min at 4° C. The cells were then washed and resuspended in Medium-199 supplemented with 2% FBS with 0.25 μg 7AAD (BD Biosciences). Flow cytometry was performed on a BD FACS Aria III sorter (BD Biosciences) and all data were analyzed using FlowJo (Treestar).
Intracellular DPP4 Staining
Bone marrow from leukemia-bearing mice was incubated with anti-mouse Fc Block (BD Biosciences, dilution 1:50) for 10 min at 4° C. followed by surface staining with anti-CD45-APC/Cy7 (BD Biosciences, dilution 1:100). Samples were washed and stained with LIVE/DEAD fixable viability dye (ThermoFisher) in accordance with the manufacturer's instructions. The cells were thereafter fixed with Cytofix/Cytoperm (BD Biosciences) for 20 min at 4° C. 1×Perm/Wash buffer (BD Biosciences) was then used to wash the cells and the cells were incubated with either anti-DPP4 or an isotype control antibody (Biolegend, dilution 1:20 for both) that were both conjugated to Alexa Fluor 647 in-house (Abcam) for 30 min at room temperature. The cells were then washed one last time in Perm/Wash buffer (BD Biosciences) and resuspended in Medium-199 supplemented with 2% FBS for analysis. Flow cytometry was performed on an LSR II instrument (BD Biosciences) and all data were analyzed using FlowJo (Treestar).
Cell Cycle Analysis
Bone marrow isolated from leukemic mice was blocked with anti-mouse Fc Block (BD Biosciences, dilution 1:50) for 10 min at 4° C. Surface staining with anti-CD45-APC/Cy7 (dilution 1:100) and DPP4-PE (dilution 1:20) was performed at 4° C. for 30 min. Following this, the samples were washed in Medium-199 supplemented with 2% FBS and then fixed with Cytofix/Cytoperm (BD Biosciences) for 20 min at 4° C. The fixed cells were thereafter washed with 1×Perm/Wash buffer (BD Biosciences) and resuspended in Perm/Wash buffer containing anti-Ki67-AF647 at a 1:10 dilution for a 30 min incubation. The samples were washed one more time with 1×Perm/Wash buffer (BD Biosciences) and then incubated in 1×Perm/Wash buffer (BD Biosciences) with 2 μg ml−1 DAPI (Biolegend) for 10 min. Finally, the samples were spun down to remove the DAPI-containing buffer and resuspended in Medium-199 supplemented with 2% FBS for analysis. Flow cytometry was performed on a BD FACS Aria III sorter (BD Biosciences) and all data were analyzed using FlowJo (Treestar).
Analysis of DPP4 Expression Following Co-Culture
A total of 250,000 MLL-AF9 or HoxA9-Meis1 leukemia cells were plated at a 1:1 ratio with the following stromal cell lines:NIH-3T3 (American Type Culture Collection, ATCC), MS-5 (RIKEN), MLO-A5 (Kerafast), MC-3T3-E1 (ATCC) and SV40 immortalized bone marrow stroma. The cells were grown in RPMI 1640 (Gibco) supplemented with 1% Penicillin-Streptomycin (Gibco), and 10% FBS (Gibco). MLL-AF9 cultures were supplemented with 20 ng ml−1 SCF, 10 ng ml−1 IL-3 and 10 ng ml−1 IL-6 (all cytokines from Peprotech). HoxA9-Meis1 cell cultures were instead grown in 100 ng ml−1 SCF and 5 ng ml−1 IL-3. The cells were co-cultured for 3 days. For flow cytometry of DPP4, the co-cultures were trypsinized and subsequently blocked with murine Fc Block (dilution 1:50) for 10 min at 4° C. Surface staining with anti-CD45-APC/Cy7 (dilution 1:100) and DPP4-PE (dilution 1:20) was then carried out for 30 min at 4° C. The cells were thereafter washed and resuspended in Medium-199 supplemented with 2% FBS with 0.25 μg 7AAD (BD Biosciences) for analysis. Flow cytometry was performed on an LSR II instrument (BD Biosciences) and all data were analyzed using FlowJo (Treestar). For the analysis of DPP4 mean fluorescence intensity, each sample was normalized to a corresponding isotype control antibody-stained sample.
Myeloid Lineage Markers
Bone marrow collected from C57B1/6J mice was stained with a hematopoietic stem and progenitor cell cocktail for 45 min at 4° C. The cells were then washed and resuspended in Medium-199 supplemented with 2% FBS with 0.25 μg 7AAD (BD Biosciences). Flow cytometry was performed on a BD FACS Aria III sorter (BD Biosciences) and all data were analyzed using FlowJo (Treestar).
Reverse Transcription with Quantitative PCR
RT-qPCR was performed to determine levels of Dpp4 mRNA. A total of 1×106 leukemia cells or T cells were lysed and the RNA was extracted using the RNeasy Plus mini kit isolation kit (Qiagen). RNA was subsequently reverse transcribed into cDNA with the SuperScript IV First-Strand Synthesis System (ThermoFisher). The qPCR analysis was performed using iTaq Universal SYBR Green Supermix (Biorad) with primers specific for Dpp4 (forward, ACCGTGGAAGGTTCTTCTGG (SEQ ID NO: 1); reverse, CACAAAGAGTAGGACTTGACCC (SEQ ID NO: 2)) and Gapdh (forward, TGTGTCCGTCGTGGATCTGA (SEQ ID NO: 3); reverse, TTGCTGTTGAAGTCGCAGGAG (SEQ ID NO: 4)). Threshold values (CT) were estimated using CFX Maestro (Biorad) and transcript levels were normalized by subtracting the corresponding Gapdh values. The relative amount of RNA is presented as 2−ΔΔCt.
Droplet-Based Single-Cell RNA Sequencing
WCBM and Image-seq samples were counted in a hemocytometer and encapsulated for a maximum output of 8,000 cells into emulsion droplets using the Chromium Controller (10× Genomics). scRNA sequencing libraries were subsequently prepared using Chromium Single Cell 3 v2 Reagent kits (10× Genomics). Reverse transcription and library preparations were done on a Biorad T100 Thermo Cycler (Biorad). cDNA libraries and final libraries were quantified on an Agilent BioAnalyzer (Agilent Technologies) using a High Sensitivity DNA kit (Agilent Technologies). Libraries were diluted to 4 nM and pooled before sequencing on the NextSeq 500 Sequencing system (Illumina). Pools were sequenced with 75 cycle run kits (26 bp Read1, 8 bp Index1 and 55 bp Read2) to a saturation level of ˜70-80%.
SMARTseq-v4 Library Preparation and Sequencing
Libraries were prepared using the MANTIS Liquid Handler (Formulatrix) and the Biomek FXP Single Arm System with Span-8 Pipettor (Beckman Coulter). Full-length cDNA was prepared using the SMARTseq-v4 Ultra Low Input RNA Kit for Sequencing (Takara Bio) and sequencing libraries prepared using the Nextera XT DNA library preparation kit (Illumina).
The SMARTseq-v4 assay utilizes the SMART technology switching mechanism at the 5′ end of the RNA template to generate full-length cDNA from as little as 10 pg total RNA. The cDNA was assessed for concentration using the Quant-iT Picogreen dsDNA assay kit (Invitrogen, P7589) on the SpectraMax i3 Multi-Mode Detection Platform (Molecular Devices) and normalized to 0.2 ng μl−1 prior to library preparation. Full-length cDNA was fragmented using the Nextera technology in which DNA is simultaneously tagged and fragmented. Tagmented samples were enriched and indexed using 18 cycles of amplification with PCR primers, which included dual 8 bp index sequences to allow for multiplexing (Nextera XT Index Kit). Excess PCR reagents were removed through magnetic bead-based cleanup using PCRClean DX beads (Aline Biosciences) on a Biomek FXP Single Arm System with Span-8 Pipettor (Beckman Coulter). The resulting libraries were assessed using a 4200 TapeStation (Agilent Technologies) and quantified using qPCR (Roche Sequencing). Libraries were pooled and sequenced on a NextSeq Mid Output flow cell using paired, 75 bp reads (Illumina).
10× scRNA-Seq Data Processing
For the 10× scRNA-Seq data, fastq files were obtained using bc12fastq (v1.8.4). Reads were aligned to the mm10 mouse reference genome using the Cellranger pipeline (v3.0.2, 10× Genomics) with default parameters. The obtained read count matrices were further filtered based on two quality metrics: the number of total UMI counts per cell (>800); and the mitochondrial transcript ratio per cell (<0.2). Conos (v1.4.1) 21 was used to integrate multiple scRNA-seq datasets. Each individual dataset was first normalized using the basicP2proc function in pagoda2 (v1.0.10) using default parameters (https://github.com/kharchenkolab/pagoda2/releases/tag/v1.0.10). Different samples were then aligned using Conos with default parameter settings (PCA space with 30 components, angular distance, mNN matching, k=15, k.self=5), and UMAP embedding was estimated using default parameter settings. Leiden clustering (conos::findCommunities) was used to determine joint cell clusters across the entire dataset collection. Quality parameters for the 10× scRNA-seq data are listed in
Differential Expression
For differential expression analysis between cell types in the 10× scRNA-seq data, a Wilcoxon rank sum test, implemented by the getDifferentialGenes( ) function from Conos R, was used to identify statistically significant marker genes that were expressed in each cell cluster. The genes were considered differentially expressed if the P value-determined Z score was greater than 3. For differential expression analysis between Image-seq and WCBM (for example, Image-seq monocytes versus WCBM monocytes), the getPerCellTypeDE( ) function in Conos was utilized.
Cell Annotation
Annotation of the cluster communities was done using marker gene expression. Initial annotations were obtained by entering the top 100 differentially expressed genes in each cluster (ordered by logFoldChange) into the CellKb database65 and further refined by consulting the primary literature referenced therein along with other relevant publications. Specificity of selected markers was additionally confirmed by evaluating expression in the Haemopedia66 and CellMarker67 bone marrow datasets. Markers to classify ten major cell types were identified: B cell (Ms4a1, Ly6d, Cd79a), pre/pro-B cell (Vpreb1, Vpreb3, Dntt), basophil (Mcpt8, Prss34, Ms4a2), dendritic cell (Bst2, Irf8, Siglech, Cox6a2), erythroblast (Cart, Hemgn, Ctse, Cpox, Atpif), monocyte (Lyz2, Ctss, s100a4), monocyte progenitor (Ly6c2, S100a10), diverse progenitors (Cd34, Prtn3, Mpo, Elane, Mpl), granulocyte progenitor (S100a8, S100a9, Cebpe, Fcnb) and neutrophil (Mmp8, Ifitm6, S100a11, S100a8, S100a9).
SMARTseq-v4 Sequencing Data Analysis
SMARTseq-v4 sequencing data were aligned with hisat2 (v4.8.2), using the genome reference mm10. FeatureCounts69 (v1.6.4) was used to calculate read counts. The quality of cells was then assessed based on the number of total counts per cell (library size), requiring at least 500,000 reads per cell. A total of 84 AML and 43 stroma cells were retained for downstream analysis. Seurat was used to analyze the SMARTseq-v4 data, and the quality parameters are listed in
Cell Cycle Signature Score
To assess cell states in different cell subsets and conditions a gene set signature score was used to measure the relative difference of cell cycle states. The signature scores were calculated as average expression values of the genes in a given set. The signature gene list was downloaded from Whitfield et al.70. Hierarchical clustering of cell cycle signature score was used to group AML cells. A two-sided Student's t-test was used to assess differential expression of selected cell cycle genes between proliferating (P) and non-proliferating (NP) cells.
Regressing Out Cell Cycle Genes
Seurat71 (v4.0.6) was used to regress out cell cycle genes. First, each cell was assigned a score, based on its expression of G2/M and S phase markers with the CellCycleScoring function. Then the ScaleData function was applied to regress out the cell cycle genes. The scaled residuals of this model represent a ‘corrected’ expression matrix that can be used downstream for dimensionality reduction. UMAP embedding and graph-based shared nearest neighbor clustering were used with five principal components, and Seurat::FindClusters( ) was used to identify AML cell sub-clusters. Seurat::FindAllMarkers, which utilizes a two-sided Wilcoxon rank sum test to assess statistical significance, was used to find the differentially expressed genes within each sub-cluster.
Analysis of Differentially Expressed Genes
DESeq2 was used to analyze differentially expressed genes between P and NP cells (
GO Term Enrichment
To test for enriched GO biological processes in gene sets, the ClusterProfiler72 (v4.0.0) package was used to evaluate the enrichment of GO categories in the sets of upregulated and downregulated genes and rank them by adjusted P value (
Analysis of Human AML Data
A regular gene expression correlation analysis was applied to two published, bulk RNA-seq datasets (531 patients for the Oregon Health and Science University, OHSU dataset48, and 188 patients for The Cancer Genome Atlas, Firehose dataset49). Spearman correlation coefficients for each gene with DPP4 were calculated. The top 300 positively correlated genes (based on strength of the correlation coefficient) were determined for each of the two datasets (Firehose and OHSU). Interestingly, a high degree of overlap was observed between the top 300 positively correlated genes from the two datasets. The overlapping gene set was then used for GO analysis.
Multiphoton Imaging Experiments
Image Acquisition
To image the AML cell distribution in different cavities, as well as CellTracker CMTPX retention, two-photon excitation at 900 nm was used, and the emission was collected at 340-460 nm to detect the second-harmonic generation signal of collagen (bone), while 500-550 nm was used to detect the AML-GFP signal. The bone front staining and CMTPX were excited at 775 nm and the resulting fluorescence collected using 525/50 nm (tetracycline) and 617/73 nm (Alizarin Red, CMTPX) filters. All image stacks were acquired using a previously described microscope51,61, with a 2 μm step size from the calvaria surface, and 20 frames from the live scanning microscope (30 frames per second) were averaged to acquire a single image.
For imaging the stroma, β-actin-GFP mice, DPP4 expression in vivo, as well as for quantifying AML proliferation, the home-built microscope described above was used, with an imaging wavelength of 980 nm. Emission signals were collected with filter 439/154 for the second-harmonic generation signal, filter 525/50 for GFP and filter 607/70 for the AF568 signal with a combination of FF705 LP, FF495 LP and FF552LP dichroic mirrors. For imaging the CXCL12-DsRed stroma, two-photon excitation at 980 nm was used, and emission signals were collected using the filter 439/154 for the second-harmonic generation signal (blue channel, shown as gray in the Figures) and 585/40 for DsRed (red channel) with the same dichroic mirrors as listed above. Image stacks were acquired with a step size of 2 μm, as well as a 15-frame average.
For displaying the data, some images were background-subtracted with the mode of the stack histogram (corresponding to the noise pixel intensity) and subsequently filtered using the 3D Fast Filters (median) function in FIJI with an x, y and z radius unit of 1. The brightness and contrast of images in the figures were adjusted, but in all cases the image analysis was performed using the raw data.
Image Stitching and Maximum Intensity Projections
Large area images were obtained by stitching together images from individual microscope fields of view sequentially for each z plane using the Grid/Collection stitching plugin in Fiji and using an overlap of 30%. Maximum intensity projection images were obtained using the Z Project function (Fiji).
Characterization of P, NP and IM Cells
The same bone marrow cavities in the same animals were imaged both on day 1 and day 3 after transplantation of AML cells. Identical cavities on day 3 were found using the recorded coordinates with respect to the bregma and lambda reference points, and comparison of the signal distribution and specific landmarks in the second-harmonic generation channel. The number of AML cells was quantified at both timepoints, and the fold-change in AML cells between day 1 and day 3 was calculated for each cavity:Fold-change=(no. of cellsday 3−no. of cellsday 1)/(no. of cellsday 1).
Cells were then grouped based on their fold-change as either proliferating (P, fold-change >2), intermediate (IM, fold-change>0 and ≤2), or non-proliferating (NP, fold-change≤0), which corresponded to average cell numbers of 29.6 (with a 99.9% confidence interval of 21.7 to 37.4 cells), 13.2 (with a 99.9% confidence interval of 5.5 to 20.8 cells), and 2.1 (with a 99.9% confidence interval of −1.2 to 5.4 cells), respectively, on day 3.
Analysis of Cell Tracker Labeling
AML cells were labeled with CellTracker Red (CMTPX, 10 μM, ThermoFisher Scientific) before transplantation. In brief, the cell suspension (in Ca2+/Mg2+ free PBS containing 10 μM CMTPX) was incubated at 37° C. for 45 min. Cells were then pelleted to remove the staining solution and resuspended in 300 μl PBS for retro-orbital injection. The mean CellTracker intensity was measured on day 3 after transplantation and measured at the brightest plane of the cell. The cells were considered positive for CMTPX when the measured signal was greater than 12.5 (the background noise in the marrow cavity). The number of CMTPX-positive cells was then divided by the total cell number sampled from the bone marrow cavities harboring the same cell counts (n=7 mice).
Analysis of DPP4 Expression In Vivo
DPP4 antibody (BioLegend) and Isotype control antibody (BioLegend) were conjugated to AF568 using the Lightning Link kit (Abcam). Antibody and isotype were injected retro-orbitally 1 h prior to the imaging session at a dosage of 1 mg kg−1. Images were collected with an excitation wavelength of 980 nm (1.6 nJ pulse energy, filter set detailed above), using a z-step of 2 μm with a 15-frame average on the custom-built microscope described above.
AML cells were segmented based on the GFP signal in the obtained images. For this, seeds were generated using the interactive watershed tool (Fiji) and used as input for the 3D-Watershed segmentation plugin (Fiji). The images were thresholded to generate a mask. This was used to calculate the total AF568 signal, as well as the total GFP signal, in each cell using the red and green channels of the images, respectively, as input for the 3D Object counter plugin (Fiji). The ratio of red to green fluorescence (multiplied by 10) for each cell was plotted in
Distribution of AML Cells in D-, M- and R-Type Cavities
The protocols to determine the bone remodeling status have been described previously. The first calcium-chelating reagent dye 1 (Tetracycline, Sigma, 35 mg kg−1) was given i.p. 48 h prior to imaging to label the bone fronts and track the bone resorption activities over the course of 2 days. The second calcium-chelating reagent dye 2 (Alizarin Red, 40 mg kg−1) was injected on the day of imaging to label all of the bone fronts. The bone remodeling status was then defined based on the ratio of dye 1 to dye 2 in a single bone marrow cavity (the concave endosteum), and therefore the bone marrow cavities were classified as: deposition type (D-type; dye 1:dye 2 ratio>75%); resorption type (R-type; dye 1:dye 2 ratio<25%), or mixed type (M-type; dye 1:dye 2 ratio 25-75%. The distributions of seeding and expansion of AML cells were then mapped with respect to the D-, M- and R-type cavities on day 0 (3 h after transplantation), day 1 and day 3 after transplantation. The same mouse was followed up longitudinally on day 0 and day 1 and the cavity type was defined on day 0. A separate cohort of animals was used for day 3 to avoid unwanted inflammatory responses from the survival surgical procedures.
Animal Handling
Male and female 8-week-old C57B16/J mice (cat. no. 000664) or male and female CXCL12-DsRed mice (cat. no. 022458) were ordered from the Jackson Laboratory, housed in an animal facility for at least 2 weeks and used for experiments at between 10 and 14 weeks of age. The 8-week-old female Fezh/J mice B6J.129(Cg)-Gt(ROSA)26Sortm1.1(CAG-cas9*,−EGFP)Fezh/J were ordered from the Jackson Laboratory (cat. no. 026179) and used in experiments. Male and female β-actin-GFP mice (Jackson Laboratory, cat. no. 006567) were bred in-house and used between 10 and 16 weeks of age. Male and female β-actin-DsRed mice (Jackson Laboratory, cat. no. 006051) were bred in-house and used between 10 and 16 weeks of age. The β-actin luciferase (βact) mice from Taconic (cat. no. 11977) were bred with ubiquitin-c-GFP (UcGFP) mice from the Jackson Laboratory (cat. no. 004353) to generate βact-UcGFP transgenic mice. An 8-week-old female βact-UcGFP mouse was then used to generate the HA9M1 cell line. All mice were housed in the pathogen-free Massachusetts General Hospital (MGH) Animal Facilities, which were equipped with ventilated micro-isolator cages. Sentinel programs and veterinary oversight were in place. Mice were given standard chow and drinking water ad libitum. An automated 12 h dark-12 h light cycle was observed and mice were housed at a fixed temperature (21° C.) and humidity (66%). The MGH Animal Facility is under the supervision of the MGH Center for Comparative Medicine. All facilities are fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (000809) and meet the National Institutes of Health standards as set forth in the Guide for Care and Use of Laboratory Animals by the Department of Health and Human Services. All procedures involving animals were carried out in agreement with protocols 2012N000190, 2007N000148 or 2016N000085 approved by the Institutional Animal Care and Use Committee of Massachusetts General Hospital.
Statistics and Reproducibility
P<0.05 was considered significant unless specified otherwise.
System Requirements
Imaging Modality
Either a confocal or multiphoton microscope can be used to visualize the 3D spatial organization of bone marrow (BM) tissue.
Ablation Modality
Requires a scanning device that steers the laser ablation beam across the microscope field of view. This has similar characteristics as the scanning optics used in some multiphoton microscopes (in fact, in this system, the same scanning optics are used for imaging and laser ablation). The ablation laser should fulfill the following requirements: repetition frequency between 100 kHz to 5 MHz (the latter is used on our system) and a pulse energy of ˜50-100 nJ. Note that while we typically use between 10-20 nJ pulse energy for the Image-seq procedure, the laser itself should have higher pulse energies because typical optical systems incur energy losses as the beam propagates to the image plane of the microscope. If the ablation geometry can't be controlled by arbitrarily positioning the scanning angle of the device, a variable aperture should be inserted into the intermediate image plane to control the ablation geometry.
Flushing System
A flushing system should be installed on the objective lens of the multiphoton/confocal microscope, as shown in
Micropipette Assembly for Micropipette
The micropipette holder should be mounted onto a motorized micromanipulator (Sutter Instrument, MPC385), and connected to an air syringe (Cooper Surgical, 6-34-520) with polythene tubing (Cooper Surgical, 6-34-536). To aid with quickly translating the micropipette to the bone marrow and pulling it back out to eject the collected sample into a tube, the micromanipulator should be mounted onto a sliding stage. A sliding stage can be made by removing the actuator from a translational stage and sliding in a post (L=50 mm) as a spacer, thereby achieving an IN (post in) or OUT (post out) position (see
Visualizing the Tissue for Image-Seq
Multiple contrast mechanisms can be used to visualize the procedure and reconstruct the 3D spatial position of the extracted cell sample, including autofluorescence, confocal reflectance or labeling with fluorescent membrane dyes such as DiD/DiR/DiI/Di8-ANEPPS or a fluorescent antibody. If a single cell or a specific cell population is to be isolated, then either transgenic expression or fluorescent antibody labeling of this subpopulation is necessary so that after aspiration of the target cell(s) and 100-400 surrounding cells, the target cells can be isolated by flow cytometric sorting. For transplantation models, cells can be labeled prior to injection. If not working with a fluorescent reporter mouse or using confocal reflectance to visualize the tissue, we recommend injecting Di8-ANEPPS (1.9 mg/kg animal weight) prior to the Image-seq procedure.
IACUC Approval
Prior to performing any of the experiments detailed below, an animal protocol describing these procedures should be submitted to and approved by the institutional animal care and use committee (IACUC). Experiments should be carried out in accordance with the guidelines set forth for the procedures on rodents at the home institution. Administer analgesics in accordance with an institution's IACUC policy and consult an OAR/IACUC veterinarian.
In Situ Imaging of Long Bones
Materials:
A. Euthanize the animal and dissect the tibia (or femur bone, or any other type of bone) from the animal. Use Kimtech wipes to remove excess muscle tissue from the isolated bone.
B. Carefully scrape off superficial bone tissue from the bone surface with a razor blade (or a micro drill) until the bone marrow becomes visible (see
C. Mount bone sample on a microscope slide by gently pressing the bone into the modelling clay (blue in
In Vivo Imaging of the Calvarium
Materials:
Procedure
1. Prior to beginning the experiment, install the correct dichroic mirrors and filters for visualizing and separating the emission spectra of the fluorophores of interest. Tools like ThermoFisher's Fluorescence spectra viewer or BioLegend's spectra analyzer can be used to find optimal choices.
2. Administer analgesics ˜30-60 min prior to imaging (we use buprenorphine IP at 0.05-0.1 mg/kg animal weight). Anesthetize the mouse (we use vaporized isoflurane at 3-4% for induction and 1-2% for maintenance), ensuring depth of anesthesia by toe pinch. Hydrate the mouse's eyes.
3. Transfer the mouse to a holder with integrated heating pad and anesthesia supply.
4. Trim the hair around the incision site using a sterile razor blade or surgical scissors and scrub with betadine.
5. Make a ˜5 mm×7 mm incision into the skin (see
6. Fold back the skin flap, secure it to the back of the skull using a drop of antibiotic ointment and expose the skull bone. Hydrate it with a drop of sterile PBS, taking care to keep the skin flap hydrated as well.
7. Remove the periosteum by gently rubbing two sterile cotton swabs across the skull in a concerted motion, starting from the interfrontal/sagittal suture and moving the swabs outwards.
8. Transfer the mouse to the multiphoton/confocal microscope. Secure in such a manner that the skull surface is mounted parallel to the image plane of the microscope.
9. Record the position of lambda and bregma (see
10. Perform IVM experiments and record the position of the imaging site with respect to Bregma and Lambda. We often record z-stacks with a 2 μm step size and 15-30 frame average. To study short-term dynamics, it is helpful to perform time-lapse microscopy and record images or short z-stacks with a time interval of a few seconds or minutes, or up to a few hours. If this is a terminal experiment, sacrifice the animal after finishing the imaging session.
For Survival Experiments:
11. Once the imaging session is completed, flush the skull extensively with sterile saline. Close the scalp with surgical sutures and infiltrate 0.25% Bupivacaine (2 mg/kg) into the surgical site to aid with pain management. Finally, apply triple antibiotic ointment (Curad) on the sutured area. Return the animal to its cage and monitor it until awake. Administer buprenorphine (0.05-0.1 mg/kg animal weight) either IP or SQ along with topical antibiotic ointment every 8-12 hours for up to two days after the surgery.
Image-Seq Procedure
Equipment:
Materials:
Permanent Marker
Prior to Executing the Image-Seq Cell Isolation
A. Coat the micropipette to prevent cells from sticking to the glass surface of the micropipette it is necessary to coat it with a siliconizing agent that prevents protein adhesion. It is also recommended to fluorescently coat the pipette for visualization by multiphoton/confocal imaging. This can be achieved as follows:
Install a fresh set of tubing onto both channels of the peristaltic pump described above, attaching the micropipette to one end. Hold the micropipette into a Falcon tube filled with 1-5 mL of Sigmacote. Allow the Sigmacote to flow into the micropipette through its tip, setting the flow rate to ˜200 μl/min, and allowing it to flow continuously for 2 min. The waste can be collected by inserting the other end of the tubing into an empty Falcon tube.
Transfer the micropipette to the other channel and flush it with sterile, deionized (RNAse-free) water using the same flow speed and direction. We typically coat many micropipettes at once and store them for future use.
Prior to experiment, install the micropipette on multiphoton microscope, and flow a fluorescent dextran with a color of choice into the micropipette using the air syringe. Pipette the dextran up and down for 2-3 min before expelling the liquid and gently drying the tip with a piece of lens tissue. The micropipette is now ready for use. Note that we do this final step directly before the experiment.
B. Prepare microfuge tubes for ejecting the samples: Clean the workspace with RNAse Away. Pipette 5 μl of Medium-199 supplemented with 2% FBS into enough microfuge tubes to collect all micropipette samples. Store them on ice. Clean the microfuge holder (
C. Visualization: If not working with a fluorescent reporter mouse/fluorescent transplant model, it is necessary to inject either Di8-ANEPPS (1.9 mg/kg animal weight) or a fluorescent antibody to aid in the visualization of the Image-seq procedure and/or label the target cell population. Wait at least 30 min after injection to ensure proper labeling.
D. Perfusion: Unless doing in vivo cell isolation, mice should be perfused prior to sample isolation. This minimizes the RBC content and further processing steps. General procedures for rodent perfusion have been described in detail elsewhere, along with relevant videos. We detail our adaption below:
Prepare 40 ml of 5 mM EDTA/PBS in a Falcon tube, along with 40 ml of regular PBS (w/o Ca/Mg).
Store both solutions on ice.
Set the flow rate of the peristaltic pump to 5 ml/min and attach a 25G needle to one end of the tubing
4. Use a permanent marker to mark the needle ˜5 mm from its tip (note that the exact distance depends on the size, and therefore the age and gender, of the mouse heart). This is to ensure that the needle is not injected too deeply into the mouse heart during the perfusion procedure.
Transfer mouse to a dissection tray and set the anesthesia to 4-5% vaporized isoflurane.
Shave the chest.
Place one end of the peristaltic tubing into the solution of EDTA/PBS, turn on the pump and wait until liquid is flowing out of the syringe. Turn off the pump.
Scrub the chest with betadine, make an incision and cut off the skin in the region surrounding the ribcage (
Cut open the ribcage parallel to the sternum, at a distance of ˜10 mm on either side. Fold up the ribcage towards the mouse head and gently move the beating heart with a pair of tweezers to expose the right atrium. Make an incision into the right atrium and ensure that blood is flowing into the chest cavity.
Use a pair of tweezers to gently reposition the heart and facilitate insertion of the needle into the apex of the left ventricle. Turn on the pump and hold the needle in place with a pair of tweezers (
Perfuse with 10 ml of ice-cold PBS/EDTA followed by 10 ml of ice-cold PBS.
Mount the mouse in a holder or dissect the calvarium/tibia and mount it onto a glass slide as described in the section for in situ imaging of long bones described above.
E. Perfusion for the isolation of stromal cells: This procedure is nearly identical to the one described above. However, instead of perfusing with ice cold EDTA/PBS, the mouse is first perfused with 10 ml EDTA/PBS at 37° C., then with 10 ml of an enzymatic digestion buffer at 37° C. This buffer consists of 450 U/mL Collagenase I (Sigma), 125 U/ml Collagenase XI (Sigma), 60 U/ml Hyaluronidase (Sigma), 60 U/ml DNase I (Sigma) in 20 ml of Medium199 (Gibco). The mouse is then incubated at 37° C. for 20 min and the calvarium or long bone dissected and mounted.
Laser Ablation and Cell Aspiration by Micropipette
After positioning the sample (this can be either a live mouse or a mounted calvarium/long bone) in the image plane of the microscope and finding the target location, turn on the peristaltic pump and flow PBS across the sample at a flow rate of 10 ml/min.
A. Check the thickness of the bone by translating the sample along the z-dimension, using the second harmonic generation signal (SHG signal, collected at half the wavelength of the multiphoton laser) to visualize the bone.
B. Set the laser pulse energy to 14 nJ at the sample and thin a 200×300 μm area of bone, leaving a layer that has a minimum thickness of 20 μm, thereby ensuring minimal damage to the bone marrow located beneath (see
C. Set the laser pulse energy to 11 nJ and ablate a ˜60×100 μm microchannel (
D. Move the sample down (z-dimension) and slide in the micropipette (“IN” position), visualizing its tip in the microscope field-of-view.
E. Move the sample (=mouse in
F. Aspirate the target cells while recording a video by gently suctioning with the air syringe (
G. Move the sample down and slide out the micropipette (“OUT” position).
H. Place a microfuge tube filled with Medium-199/FBS onto a holder and insert the micropipette into the tube liquid. Expel the aspirated cell sample, directly generating a single-cell suspension (see
I. Remove the tube and place it on ice.
J. Repeat the procedure to collect more Image-seq samples. To ensure high-quality single-cell data, ensure that the entire procedure (from mouse perfusion to multiple sample pickings) does not take longer than ˜2.5 h. In this time-frame we typically collect 4-6 samples.
In general, the number of Image-seq samples that are required to obtain statistical significance is given by the size of the biological effect and the observed standard deviation. If there is no sequencing data available in the literature to estimate the size of the biological effect, we recommend running a small pilot with N=3 Image-seq samples to quantify it along with the standard deviation, and, if necessary, to collect additional samples until statistical significance is achieved. Ideally samples should be collected from at least 3 different mice. Depending on the biological question, use either the 10× genomics platform for cell encapsulation and library preparation or the SMARTseq-v4 protocol
10× Procedure
Materials:
A. All of the droplet-based sequencing data generated for the Image-seq paper used 10× Genomics v2 kits that are no longer available for purchase. Depending on the sequencing needs consult with local sequencing core or 10× Genomics representative regarding what type of 10× Chromium Single Cell analysis kit needed.
B. Check the final volume in each microfuge tube from the Image-seq cell isolation procedure described above. At this stage there are two considerations. Firstly, the sample should be in a volume that is compatible with the 10× Chromium Single-Cell analysis kit of choice. Secondly, the overall cell number needs to be determined. For counting cells in the disposable hemocytometer, the minimal required volume is 5 μl. Adjust the volume of the samples so that 5 μl can be safely removed without depleting the sample for the downstream sequencing step. If sample size is not a limiting factor, a larger volume of the sample can be used to perform the cell count.
C. Mix the cell suspension in a 1:1 ratio with 0.4% trypan blue.
D. Load the cells onto the hemocytometer (iNCYTO) and count live and dead cells. Also, take note if there are cell aggregates. If there is extensive aggregation, additional dissociation of the sample could be required to avoid clogging the 10× Chromium Chip.
E. Consult the cell loading table in the Chromium Single Cell 3′ user guide. Using 0.4 mg/ml ultra-pure BSA in Medium 199 adjust the cell concentration accordingly.
F. Place the cells on ice until ready to load on the 10× Chromium Chip.
G. For preparation of the scRNA-seq samples, follow the instructions provided by 10× Genomics kit.
H. In order to prepare scRNA-seq libraries for sequencing, consult with sequencing core or provider to determine preferred submission format.
SMARTseq-v4 Procedure
If SMARTseq-v4 library preparation is chosen for processing the samples, it is necessary to first sort single cells into individual wells of a 96-well plate using a flow cytometer or microfluidic sorting device. To aid with drawing the correct gates, it is recommended to collect whole bone marrow preparations from one of the bones that were not processed by Image-seq. For our experiments, the Bauer Core at Harvard University performed library preparation and sequencing using the SMARTseq-v4 Ultra Low Input RNA kit. It is recommended that using a core facility or a company to perform library preparation and sequencing since the procedure can be time-consuming and tedious.
Equipment:
Flow Cytometer
Any equipment detailed in the SMART-Seq® v4 Ultra® Low Input RNA Kit for Sequencing User Manual.
Materials:
Any further reagents listed in the SMART-Seq® v4 Ultra® Low Input RNA Kit for Sequencing User Manual.
Sort
A. Pipette sample up and down several times to generate a single-cell suspension.
B. Stain sample with antibody for 30 minutes at 4° C. in Medium 199 supplemented with 2% FBS, using the antibody concentration recommended by its manufacturer.
C. Add 1 ml of PBS to the sample, as well as 0.1 μg of DAPI and incubate the sample for 10 min.
D. Gently vortex the cell sample and transfer it to the flow cytometer (MoFlo Astrios EQ cell sorter for our experiments).
E. Use a whole bone marrow preparation processed in the same manner to draw initial gates for flow cytometric sorting.
F. Sort single, live cells into individual wells of a 96-well PCR plate filled with 2.6 μl of lysis buffer (Takara Bio USA, Inc.).
G. After completing the sort, seal plates, spin them down and snap freeze them
H. Store plates at −80° C. prior to preparation for cDNA synthesis using the SMARTseq-v4 assay. Note that it is possible to bank samples until a sufficient number of cells have been collected and perform library preparation on all of the collected cells simultaneously.
Library Preparation
A detailed description of the cell-lysis, reverse transcription, library preparation and indexing steps can be found in the SMART-Seq® v4 Ultra® Low Input RNA Kit for Sequencing User Manual by Takara Bio.
Sequencing
In order to prepare scRNAseq libraries for sequencing, consult with sequencing core or provider to determine preferred submission format.
Data Analysis
The 10× scRNA-seq data and SMARTseq-v4 data described in the Image-seq manuscript have been deposited into the Gene Expression Omnibus (GEO) database under GSE188902.
Operating System
This protocol assumes users have a Unix-like operating system (i.e., Linux or MacOS X), with a bash shell or similar. All commands given here are meant to be run in a terminal window. The R introductory message will start with a ‘>’ prompt.
Align the scRNA-Seq Reads to the Genome
For the 10× scRNA-Seq data, Reads were aligned to the mm10 reference genome using the Cellranger pipeline (version 3.0.2, 10× Genomics). For SMARTseq-v4 data, reads were aligned with hisat219 and featureCounts20 was used to calculate read counts.
A. Map the 10× scRNA-seq reads for each sample to the reference genome and read counts quantification:
B. Map the SMARTseq-v4 sequencing data for each cell to the reference genome and read counts quantification:
We used Conos21 (1.4.6) to integrate multiple scRNA-seq datasets together. Each individual dataset was first normalized using basicP2proc function in pagoda222 with default parameters. Different samples were then aligned using Conos with default parameter settings.
Regressing out cell cycle genes Seurat (version 3) was used to regress out cell cycle genes. First, we assigned each cell a score, based on its expression of G2/M and S phase markers with the CellCycleScoring function. Then we applied the ScaleData function to regress out cell cycle genes.
DESeq2 (pubmed: 25516281) was used for analyzing differentially expressed genes between proliferating (P) and non-proliferating (NP) cells
To assess cell states in different cell subsets and conditions, we used a gene set signature score to measure the relative difference of cell cycle states. The signature scores were calculated as average expression values of the genes in a given set. The signature gene list was downloaded from Whitfield et al (Identification of Genes Periodically Expressed in the Human Cell Cycle and Their Expression in Tumors. Mol Biol Cell 13, 1977-2000 (2002)). Hierarchical clustering of the cell cycle signature score was used to group AML, cells.
The present application is based on, and claims priority to, U.S. Provisional Application Ser. No. 63/394,274, filed Aug. 1, 2022, which is incorporated herein by reference in its entirety for all purposes.
This invention was made with government support under 5R01CA194596-04, 5R01DK115577-05, P01HL142494-04, and 5R01DK123216-03 awarded by the National Institutes of Health (NIH). The government has certain rights in the invention.
| Number | Date | Country | |
|---|---|---|---|
| 63394274 | Aug 2022 | US |
