WO2023218071A1

WO2023218071A1 - Methods and substrates for immobilizing leukocytes for single-molecule fluorescence imaging

Info

Publication number: WO2023218071A1
Application number: PCT/EP2023/062841
Authority: WO
Inventors: James H. Felce; Tyler S. Ralston
Original assignee: Oxford NanoImaging Limited
Priority date: 2022-05-13
Filing date: 2023-05-12
Publication date: 2023-11-16

Abstract

The present application relates to methods for preparing leukocytes for imaging, as well as methods for imaging leukocytes prepared using such methods, methods for classifying leukocytes based on such imaging data, and imaging substrates for use in such methods. The leukocytes are immobilized on an imaging surface having attached thereto (a) a passivation reagent, for preventing non-specific binding to the imaging surface; and (b) a CD45 capture molecule (such as an anti-CD45 antibody), the CD45 capture molecule being attached to the imaging surface via a secondary capture molecule.

Description

METHODS AND SUBSTRATES FOR IMMOBILIZING LEUKOCYTES FOR SINGLEMOLECULE FLUORESCENCE IMAGING

TECHNICAL FIELD

The present application relates to methods for preparing leukocytes for imaging, as well as methods for imaging leukocytes prepared using such methods, and methods for classifying leukocytes based on such imaging. The present application also concerns substrates suitable for capturing leukocytes for imaging. The technology has particular applicability to imaging T cells (such as CAR-T cells) using single molecule localization microscopy.

BACKGROUND

Single-cell characterization methods are of substantial value to many areas of biological and biomedical research. The parameters reported by such methods vary, but typically relate to the abundance of specific molecules in or on each cell (including proteins, nucleic acids, metabolic intermediates etc.) or morphological features of the cell or internal organelles (including cell size, granularity, dendricity etc.).

A common method for such single-cell characterization is flow cytometry. Relative to other single-cell characterization methods, flow cytometry is a high throughput method to acquire data on both biomarker abundance and cell morphology. In flow cytometry, hundreds of thousands of cells may be sampled by flowing cells through detection channels on the instrument. However, in some applications flow cytometry is limited by low sensitivity and low precision with regard to the characteristics of individual cells. Typically, to detect a specific biomarker by flow cytometry, thousands of copies of a molecule may be required because each cell is only measured for a very short period of time (e.g., on the millisecond scale, depending on the flow rate). Thus, the abovebackground signal is dependent on the integrated fluorescent intensity of a large number of molecules. Due to the integration of signal over a large number of molecules, the precision for detecting variations in the fluorescent emission form individual cells is low. The low precision is further exacerbated by the highly variable number of photons emitted by a single fluorophore within the detection time. As a result, flow cytometry may be limited to detecting large changes in biomarker numbers and even a change in factor of 10, 100, or 1000 copies/cell may not be reliably reported. However, biologically significant changes may occur at levels below the flow cytometry detection threshold. For example, low levels of gene expression or moderate changes in biomarker density can indicate differences in cellular status. Relative to flow cytometry, bulk analyses may provide increased sensitivity for a more quantitative approach to detecting biomarkers. Fluorescent microscopy methods have the potential to be significantly more sensitive to the presence of cellular biomarkers due to the fact that they collect far more light from each fluorophore and can include methods to overcome the stochastic differences in number of photons emitted by each molecule. However, bulk analyses do not provide results at the single-cell level (e.g., ELISA, quantitative proteomics, etc.). By contrast, single-cell transcriptomics (e.g., scRNAseq) offer some potential to provide results at a single cell level, however single-cell transcriptomics are limited by their shallow read depth (i.e., low sensitivity to gene expression with low concentrations of mRNA), relative to bulk transcriptomic analyses and the lack of absolute correction between transcript copy numbers and protein copy numbers mean that interpretation is highly limited.

In the applicant’s own earlier patent application, WO 2022/053624, the inventors report the use of single molecule localization microscopy (SMLM) to identify clustering of fluorescent molecules within a cell, and subsequent use of this information to characterize cells. The document discusses, in particular, the application of this technique to the characterization of T cells, in particular T cells useful for CAR-T therapy. However, whilst such analyses offer significant gains in sensitivity and information content compared to flow cytometry, they also present a challenge since (unlike with flow cytometry) the cells must be immobilized on a surface for imaging, which can lead to unwanted perturbation of the cell status and membrane organization. This is especially true of T cells, where interaction with a surface can cause activation/triggering (e.g. formation of an immunological synapse), leading to a significant shift away from the cell’s “resting” state. This can be particularly problematic when it is necessary to build up statistics by imaging multiple cells, since the time required to collect images of suitably representative sample of cells means that different cells can be at different points in their activation process over the course of the experiment, which could artificially lead to the classification of multiple different populations of cells when the starting sample in fact had a single type of cells in their “resting” state. Whilst applying a fixation to the cells ahead of imaging can help to preserve the resting state of the cell, this can also disrupt imaging, since fixation locks in the complex 3D morphological structure of the T cell which can complicate imaging. Given the effort and cost involved with cellular therapies, such as CAR-T therapy, there remains a need for more accurate and reliable techniques for classifying leukocytes, as a means for improving outcomes from cellular therapies.

SUMMARY OF THE INVENTION

In view of the above, the present inventors have developed methods and apparatus to allow imaging of cells by single molecule localization microscopy which minimize or avoid unwanted perturbation of the cell, in particular allowing preservation of the arrangement of the membrane in its resting state, or “native configuration”. The methods have particular applicability to leukocytes, given their propensity to undergo changes when brought into contact with an adherent surface for imaging.

In a first aspect, the invention provides a method of preparing leukocytes for imaging, comprising:

(i) providing an imaging surface, having attached thereto:

(a) a passivation reagent, for preventing non-specific binding to the imaging surface; and

(b) a CD45 capture molecule, the CD45 capture molecule being attached to the imaging surface via a secondary capture molecule;

(ii) adding a suspension of leukocytes onto the imaging surface, and allowing the leukocytes to adhere to the imaging surface via the CD45 capture molecule; and

(iii) fixing the leukocytes by applying a fixation agent.

The present inventors have found, in particular, that capturing via CD45 (a protein tyrosine phosphatase present on the surface of all leukocytes) attached to the surface via a secondary capture molecule allows minimal disruption to the resting state of the cell, minimizing perturbation of the organization of the cell membrane up to the point of fixing.

In particular, the present inventors have found that to facilitate control over the distribution of molecules at the cell surface, a minimum separation distance must be maintained between the basal cell membrane and the imaging surface. This is because the phosphorylation status of many lymphoid surface receptors (including TCR, CAR, and costimulatory/coinhibitory receptors) is sensitive to the passive segregation of large molecules from the contact region. Many such large molecules (e.g., CD45, CD148) are highly active tyrosine phosphatases that both dephosphorylate signaling motifs in surface receptors, and also contribute to the regulation of kinases (e.g., Src kinases) that phosphorylate the same motifs. If large molecules are excluded from the contact region then surface receptors within this region will spontaneously become more heavily phosphorylated. For example, if the cell-imaging surface distance is the approximate size of the extracellular domains, approximately >20-3 Onm, the large molecules may be excluded from the contact region promoting heavy phosphorylation. As another example, when T cells are adhered to a poly-L-lysine (PLL) on a surface, this can drive significant phosphatase segregation and so T cell activation, since the extracellular domains of large tyrosine phosphatases are impacted by the proximity of the PLL-coated surface.

In the present invention, the “CD45 capture molecule” is a molecule capable of binding to CD45, and the “secondary capture molecule” is a molecule which binds (either directly, or indirectly e.g. via an intermediary molecule) to the CD45 capture molecule. Capturing cells via a CD45 capture molecule which is itself attached to the imaging surface via a secondary capture molecule helps to distance the basal membrane from the imaging surface, minimizing or even preventing the exclusion of molecules from the contact zone between cell and imaging surface. Targeting of CD45 in particular as a means to immobilize the cells also has other advantages compared to targeting of alternative surface proteins. In particular, the high prevalence of CD45 on the surface of leukocytes means that a large number of bonds can form between the cell and the imaging surface, leading not only to a strong immobilization of the cell but also causing the cell to “spread” or “flatten” onto the surface as new bond forms. This spreading smooths out the complex morphology of leukocytes, which helps to bring more molecules within the illumination/focal volume of microscopes typically used for single molecule localization microscopy (e.g. within the evanescent field used in Total Internal Reflection (TIRF) microscopy), and to minimize variation in signal intensity caused by variation of the position of fluorophores relative to the focal plane. This flattening of the cell membrane also allows more accurate estimation of the density of fluorescently-labeled molecules on the cell surface, since it can minimize artifacts caused by the influence of cell morphology (e.g. ruffling, microvilli, pseudopodia). In addition, the high prevalence of CD45 means that even if a portion of the molecules are immobilized relative to the surface, a suitable proportion of non-bound “free” CD45 remain mobile within the membrane during step (ii) so that the molecule can continue to dephosphorylate surface receptors and thereby prevent unwanted activation or other perturbation. The leukocyte may be, for example, a lymphocyte such as a T cell, B cell or Natural Killer (NK) cell. The leukocyte may be a naturally occurring cell type, or may be an engineered cell, such as a CAR-T cell. Preferably, the cell is a T cell or a CAR-T cell. The method is particularly well- suited to preparation of T cells and CAR-T cells for imaging, since the immune response of such cells is dictated by membrane characteristics, and such cells can be particularly sensitive to perturbations caused by bringing the cell into contact with a surface for imaging. The method of the present invention allows such cells to be imaged in their “resting state”, that is without activation or “triggering” of the cells.

The imaging surface is generally provided on a suitable substrate, compatible with imaging. The substrate may be, for example, a microscope slide, optical fiber, or prism. The term “microscope slide” also extends to, for example, coverslips, and situations where a microscope slide is incorporated as part of a larger structure, such as a microtiter plate (e.g. a 6, 24, 96, 384 or 1536- well microtiter plate or at least a portion thereof, preferably the bottom of a well thereof) or a microfluidic chamber (e.g. in a microfluidic chip). Preferably the substrate comprises, consists essentially of or consists of glass or an optically transparent polymer (with the imaging surface modified as per the method above). Most preferably the substrate comprises, consists essentially of or consists of glass, with the imaging surface modified as taught above.

In a preferred implementation, the CD45 capture molecule is an anti-CD45 antibody.

The antibody may be, for example, a monoclonal antibody, a polyclonal antibody, or an antibody fragment such as a F(ab’)2, F(ab)2, Fab’, Fab, variable fragment (Fv), single chain variable fragment (scFv), diabodies, linear antibodies, single-chain antibody molecules, and multispecific antibodies formed from antibody fragments. Most preferably, however, the CD45 capture molecule is an anti-CD45 antibody, preferably a monoclonal anti-CD45 antibody. Using whole antibody is preferential compared to using antibody fragments since they are straightforward to produce, increase the distance between the imaging surface and the basal membrane of the cell, and ensures strong binding due to avidity.

Preferably, the anti-CD45 antibody targets a binding epitope on CD45 located distally to the cell membrane. This helps to maximize the separation distance between cell and imaging surface. For example, the antibody may target an epitope in the upper 50% of the height of the extracellular domain, preferably in the upper 40%, more preferably in the upper 30%. For example, it is preferred that the anti-CD45 antibody (or variants mentioned above) targets an epitope within or including the first immunoglobulin (Ig) domain of CD45 (as determined from the N-terminus) or the common domain of the mucin-like domain of CD45, since these domains are positioned distal from the cell membrane.

Examples of anti-CD45 antibodies suitable for use in the invention include, for example, YAML 501.4 (a monoclonal IgCha antibody sold by Santa Cruz Biotechnology®, Inc., Texas, USA) or CD45-2B11 (a monoclonal IgGi antibody sold by Thermo Fisher Scientific®, Massachusetts, USA).

The passivation reagent may be, for example, a protein (e.g. bovine serum albumin (BSA) or human serum albumin (HSA)), a nonionic surfactant (such as polysorbate 20 (e.g. Tween®-20), Triton X-100, or a polymer such as poloxamer 407 (e.g. Pluronic™ F127)), a polymer such as polyethylene glycol (PEG) (optionally in the form of an ester), or any mixture thereof. The use of polyethylene glycol, such as PEG 205 or PEG 207, is preferred.

The interaction between the CD45 capture molecule and secondary capture molecule may be direct (i.e. with direct interaction between the CD45 capture molecule and secondary capture molecule) or indirect (i.e. with the interaction mediated by another molecule). In a preferred implementation, the CD45 capture molecule bears an anchor moiety, and the secondary capture molecule likewise bears an anchor moiety, wherein the interaction between the CD45 capture molecule and secondary capture molecule occurs via a mediating compound having multiple binding moieties suitable for binding to said anchor moieties.

For example, in a preferred implementation the CD45 capture molecule is a biotinylated anti- CD45 antibody, and the secondary capture molecule is PEG linked to biotin, which is then coated with neutravidin, which in turn captures the biotinylated CD45 capture molecule (i.e. where biotin serves as an anchor moiety, and neutravidin as the mediating compound).

In an alternative implementation, the CD45 capture molecule is an anti-CD45 antibody, the imaging surface is coated with neutravidin, and the CD45 capture molecule is a biotinylated secondary antibody which is used to capture anti-CD45 antibody. In such implementations, the skilled reader recognizes that alternative mediating compounds compatible with biotin may be used in place of neutravidin, such as avidin or streptavidin, or some other suitable variant thereof.

In a preferred implementation, the imaging surface is first coated with PEG linked to biotin, which is then coated with neutravidin, which in turn captures biotinylated anti-CD45 antibody. Advantageously, in this embodiment the PEG serves both as a means of passivating the imaging surface, but also as a means of tethering the CD45 capture molecule onto the surface at a distance sufficient to avoid segregation of CD45 from the contact point. Whilst the PEG layer represents a surface from which the cell is kept removed, it is much less rigid and less dense compared to most imaging surfaces (typically glass) and so is less likely to promote molecular segregation at the captured cell surface.

In preferred implementations, the imaging surface comprises (a) passivation reagent having no anchor moiety (e.g. non-biotinylated PEG) and (b) passivation reagent comprising an anchor moiety (e.g. biotinylated PEG), which is then coated with a mediating compound having multiple capture moieties suitable for binding to said anchor moiety, which in turn captures anti-CD45 antibody having an anchor moiety which binds to said mediating compound. In other words, a preferred implementation involves a modified passivation reagent serving as a secondary capture molecule.

Suitably, the density of CD45 capture molecules is the minimal number required to allow efficient capture. Excessive engagement of CD45 will enrich it within the contact area and thereby cause a reduction in surface protein tyrosine phosphorylation due to increased local phosphatase activity.

To achieve a suitable density of CD45 capture molecules whilst still permitting the surface to be suitably passivated, the ratio of passivation reagent comprising said anchor moiety to passivation reagent lacking said anchor moiety may be, for example, no more than 0.5:1, no more than 0.4:1, no more than 0.3 : 1 , no more than 0.2: 1 , no more than 0.1 : 1 , or no more than 0.05: 1. In other words, the amount of passivation reagent comprising an anchor moiety, as a percentage of the overall amount of passivation reagent, may be, for example, 50% or less, 40% or less, 30% or less, 20% or less, 10% or less, or 5% or less. In addition, to control the amount of CD45 capture molecules at the surface, the imaging surface may be treated with competitor molecules which attach to the surface via the secondary capture molecule (thereby blocking sites which might otherwise be occupied by CD45 capture molecules). In instances where the link between the secondary capture molecule and the CD45 capture molecule is mediated by an anchor moiety, the competitor molecule may be an alternative competitor molecule bearing such an anchor moiety. For example, in instances where the secondary capture molecule and CD45 capture molecule are biotinylated, and are attached via a mediating compound such as neutravidin, the competitor molecule may be free biotin, or an alternative biotinylated compound. Alternatively, if the secondary capture molecule is an antibody with selectivity for an anti-CD45 antibody, the competitor molecule may be an irrelevant antibody of the same isotype.

In particularly preferred implementations, the imaging surface comprises (a) non-biotinylated PEG and (b) biotinylated PEG, which is then coated with neutravidin, which in turn captures biotinylated anti-CD45 antibody.

In instances where it is desirable to study non-activated leukocytes (e.g. non-triggered T cells, such as CAR-T cells) in their resting state, the imaging surface may lack activating substances. For example, the passivation reagent, CD45 capture molecule, and molecules involved with attachment of those species to the imaging surface, may be the only species introduced/attached onto the imaging surface.

Alternatively, the present invention allows study of activated leukocytes (e.g. triggered T cells, such as CAR-T cells). Activated phenotypes may be induced via an immobile phase or solution phase, for example.

Immobile phase methods involve the deposition of immobile elements on the imaging surface that engage molecules on the leukocyte (e.g. T cell) surface and induce a response. For example, this may include affinity probes (e.g. antibodies), recombinant ligands (e.g. CD80 for CD28, PDL1 for PD1), non-native ligands (e.g., superantigens), and can be targeted against a wide range of surface proteins including activatory receptors (e.g., CAR, TCR, CD28, 0X40, 4-1BB etc.), inhibitory receptors (e.g., PD1, Tim3, LAG3, HVEM, CTLA4 etc.), adhesion molecules (e.g., LFA1, CD2 etc.), and/or coreceptors (e.g., CD4, CD8), antigen-presenting molecules (e.g., MHCI, MHCII, CD1). These molecules can be immobilized using the same methods as described herein for CD45 capture molecules, and can be combined in any number of ways to generate the desired phenotype.

Solution phase methods involve the capture of the leukocytes in the manner taught above using a CD45 capture molecule, and then adding to the solution soluble effector molecules to affect cell status. Such effectors include soluble ligands (e.g., pMHC, CD80, CAR ligand etc.), anti-receptor antibodies, cytokines, chemokines, toxins, cytotoxic effectors, small molecule modulators (e.g., enzyme inhibitors etc.), and chemical effectors (e.g., reducing agents, pH changes). Equivalent molecules can also be used in aggregated formats (e.g., tetramerized through streptavidin linkers) or on solid supports (e.g., polystyrene/silica beads).

In a particularly preferred embodiment, step (ii) is carried out with cells suspended in growth medium, preferably complete growth medium (containing serum). Advantageously, the combination of capturing cells via the CD45 capture strategy taught above whilst the cells are present in growth medium (instead of, for example, in an imaging buffer) minimizes perturbation of the cells prior to fixing, preserving the “native” configuration of the cell membrane as much as feasible. Furthermore, the provision of a passivation reagent on the imaging surface prevents non-specific binding of components of the growth medium (proteins or other molecules in the solution), which might otherwise interfere with imaging.

Preferably step (ii) is followed by step (ii-A), comprising flushing the imaging surface with a flushing liquid. Advantageously, flushing the imaging surface with a flushing liquid can help to remove non-adhered components from the imaging surface. The flushing liquid may be, for example, growth medium. Preferably, the growth medium lacks serum, since the serum can potentially bind to the cells. Suitably, the growth medium lacks additional cells.

The time between the start of step (ii) and the initiation of step (iii) may be, for example, at least 2 minutes, at least 3 minutes, at least 4 minutes, at least 5 minutes, at least 10 minutes, or at least 15 minutes. A period of at least 5 minutes is preferred. Advantageously, allowing the cells to adhere for several minutes increases the amount of “spreading” or “flattening” of the basal cell surface prior to fixation. In instances where the process includes step (ii-A), this step may be initiated at least 30 seconds after the start of step (ii), at least 1 minute after the start of step (ii), at least 2 minutes after the start of step (ii). Optionally, step (ii-A) is carried out shortly after the initiation of step (ii). Preferably, however, step (ii-A) is carried out just before step (iii), e.g. less than 1 minute or less than 30 seconds. For example, the time between the start of step (ii) and the initiation of step (iii) may be at last 5 minutes, with step (ii- A) carried out less than 1 minute, or less than 30 seconds from the initiation of step (iii)

During step (ii) the leukocyte may be imaged using a microscope, e.g. under white-light illumination. This imaging may show spreading of the cell as it contacts the imaging surface. Optionally, step (ii) is continued until imaging of the cell shows no or minimal further spreading.

The fixation agent used in step (iii) may be, for example, formaldehyde, glutaraldehyde, or glyoxal. 4% paraformaldehyde in phosphate buffered saline (PBS) for 10 min at room temperature is sufficient to give robust fixation without over-fixing that would damage morphology and/or epitopes required for antibody recognition.

Preferably, the fixation step is proceeded by a quenching step (iii-A). The quenching step can help to quench unreacted groups arising from the fixation (e.g. unreacted aldehyde groups). The quenching step may involve treating the leukocytes with a quenching solution, such as glycine in a suitable buffer, such as glycine in PBS (e.g. 100 mM glycine in PBS).

Optionally, the fixation step is proceeded by a blocking step (iii-B). The imaging surface is already passivated through use of the passivation reagent, but blocking can be used to prevent off-target binding to the fixed leukocytes by fluorescent stains. The blocking step may involve treating the leukocytes with a blocking solution. The blocking solution may comprise, for example, BSA, casein, serum, or other such blocking agents.

Steps (iii-A) and (iii-B) may be carried out in any order, although it is preferred that step (iii-A) precedes step (iii-B).

Suitably, the method comprises an additional step (iv) of staining/labelling the leukocytes with one or more fluorescent stains or fluorescent probes. Typically, the step (iv) is carried out after step (iii), so that the staining does not perturb the normal configuration of the cell prior to fixing. Suitable targets may include, for example, those listed in the “CAR-T specific targets” section indicated below. Optionally, the method comprises an additional step (v) of carrying out an additional fixing step after staining. This fixation step can be used to preserve staining integrity and minimize molecular motion. The fixative may be the same as those indicated above or below for step (iii).

In a second aspect, the invention provides a method of imaging leukocytes, comprising preparing leukocytes for imaging according to the first aspect, and carrying out imaging of the leukocytes. In particular, this aspect preferably comprises carrying out fluorescence imaging of leukocytes, comprising preparing leukocytes in the manner described above in relation to the first aspect and carrying out fluorescence imaging of biomarkers on the leukocytes. Preferably, the fluorescence imaging is single molecule localization microscopy (SMLM) to obtain spatial coordinates of said biomarkers. The fluorescence imaging may be, for example, dSTORM, PALM or PAINT, as described below.

In a third aspect, the invention provides a substrate having an imaging surface suitable for use in the methods taught above. More specifically, the invention provides a substrate comprising an imaging surface having attached thereto:

(b) a CD45 capture molecule, the CD45 capture molecule being attached to the imaging surface via a secondary capture molecule.

The substrate, imaging surface and passivation reagent and CD45 capture molecule may have any of the optional and preferred features discussed above in relation to the first aspect.

In a fourth aspect, the invention provides a method of making a substrate having an imaging surface as taught above. The method may comprise:

(I) treating the substrate with a functionalizing chemical to add functional groups to the surface of the substrate;

(II) adding a passivation reagent comprising a reactive group which reacts with the said functional groups to bond the passivation reagent to the substrate;

(III) adding a secondary capture molecule comprising a reactive group which reacts with the said functional groups to bond the secondary capture molecule to the substrate; and

(IV) attaching a CD45 capture molecule to the secondary capture molecule. Steps (II) and (III) may be carried out simultaneously (e.g. with a mixture of passivation reagent and secondary capture molecule), or may be carried out sequentially, e.g. Step (II) and then Step (III), or Step (III) and then Step (II).

Suitably, the functionalizing chemical is an aminosilane and the reactive group is N- hydroxysuccinimide (NHS) or an ester thereof. The aminosilane may be, for example, 3- aminopropy Itri ethoxy silane .

Preferably, the passivation reagent is PEG, and the biotinylated passivation reagent is biotinylated PEG. In especially preferred embodiments, the passivation reagent is NHS-PEG, and the biotinylated passivation reagent is biotinylated NHS-PEG. The ratio of biotinylated-PEG to nonbiotinylated PEG may be no more than 0.5:1, no more than 0.4:1, no more than 0.3: 1, no more than 0.2:1, no more than 0.1:1, or no more than 0.05:1. In other words, the amount of biotinylated-PEG as a percentage of the overall amount of PEG passivation reagent may be, for example, 50% or less, 40% or less, 30% or less, 20% or less, 10% or less, or 5% or less.

In a fifth aspect, the present invention comprises a kit of parts, suitable for making a substrate of the invention. The kit of parts may include, for example:

- said passivation reagent, being non-biotinylated PEG;

- said secondary capture molecule, being biotinylated PEG

- biotinylated anti-CD45 antibody; and

- neutravidin.

In another aspect, the present invention comprises methods of classifying leukocytes which have been imaged according to the second aspect of the invention. For example, the present invention includes a method of identifying whether a sample of T cells from a patient is suitable for use as therapeutic cells in CAR-T cell therapy, comprising: imaging the sample of T cells by single molecule localization microscopy using a method of the second aspect of the invention, to obtain spatial coordinates of a biomarker on the T cells; detecting boundaries of the plurality of cells; constructing a sample feature vector based on the obtained spatial coordinates and the detected boundaries; providing reference data, wherein the reference data comprises one or more reference feature vectors obtained for reference cells, wherein the reference cells are CAR-T cells from patients with a known therapeutic outcome; and carrying out data analysis, comprising comparing the sample feature vector with said reference feature vector(s), and determining the similarity of the plurality of cells to the reference cells, wherein a greater degree of similarity is indicative of a greater suitability for use in CAR-T cell therapy.

The following statements also set out various alternative aspects of the invention:

A method, apparatus, and system for measuring single molecule emission from a cell, the method comprising: applying a fixation agent to the cell while the cell is in a native configuration to produce a fixed cell; capturing the fixed cell on a surface of a sample substrate; and acquiring an imaging sequence of single molecule emission from the fixed cell.

Some embodiments relate to, a method, apparatus, or system to capture the fixed cell on the surface of the sample substrate comprises flattening a cell membrane of the fixed cell on the surface.

Some embodiments relate to, a method, apparatus, or system, further comprising: determining, using at least one first image from the imaging sequence, an area corresponding to the fixed cell membrane; and determining, using at least one second image from the imaging sequence, a plurality of proteins in the cell membrane and/or in the fixed cell.

Some embodiments relate to, a method, apparatus, or system, wherein the at least one first image and the at least one second image are fluorescence images.

Some embodiments relate to, a method, apparatus, or system, wherein the at least one first image is a white light image and the at least one second image is a fluorescence image.

Some embodiments relate to, a method, apparatus, or system, further comprising determining, using the at least one second image, a location of an individual emitter associated with a protein on or in the fixed cell, wherein the location is determined to within a 100 nm resolution. Some embodiments relate to, a method, apparatus, or system, further comprising determining, using the location of the individual emitter, the spatial organization of a molecule associated with the individual emitter relative to the plurality of proteins in the cell membrane and/or in the fixed cell.

Some embodiments relate to, a method, apparatus, or system, wherein determining the spatial organization of a molecule associated with the individual emitter, comprises determining the spatial organization of the number of the cell structures using a hierarchical clustering of a plurality of fluorescence signals.

Some embodiments relate to, a method, apparatus, or system to capture the fixed cell comprising capturing the fixed cell using an agnostic capture surface.

Some embodiments relate to, a method, apparatus, or system, wherein capturing the fixed cell comprises capturing the fixed cell using a capture surface configured to bind to specific cells and/or proteins.

Some embodiments relate to, a method, apparatus, or system, wherein the fixed cell is captured on a front surface of the sample substrate, and wherein acquiring the imaging sequence comprises acquiring the image sequence using a microscope configured to illuminate a back side of the sample substrate at an angle such that the illumination light undergoes total internal reflection.

The foregoing summary is not intended to be limiting. Moreover, various aspects of the present disclosure may be implemented alone or in combination with other aspects. Further, the features described in connection with one exemplary embodiment may be incorporated in other embodiments.

BRIEF DESCRIPTION OF FIGURES

In the drawings, each identical or nearly identical component that is illustrated in various figures is represented by a like reference character. For purposes of clarity, not every component may be labeled in every drawing. The drawings are not necessarily drawn to scale, with emphasis instead being place on illustrating various aspects of the techniques and devices described herein. In the drawings:

FIG. 1 A illustrates an exemplary collection of cells at the microscope scale. FIG. IB illustrates an exemplary native distribution of surface proteins at the nanoscopic scale of a cell.

FIG. 2A illustrates a method for acquiring the spatial organization data of cell structures, in accordance with some embodiments.

FIG. 2B illustrates a microscope configuration with microscope objective 216 configured to collect fluorescence emission from a cell, in accordance with some embodiments.

FIG. 2C illustrates the excitation geometry for the microscope configuration illustrated in FIG. 2B, in accordance with some embodiments.

FIG. 2D illustrates another microscope configuration compatible with imaging according to an aspect of the invention.

FIG. 3 illustrates process 300 for the initial processing and analysis of single-molecule localization data, in accordance with some embodiments.

FIG. 4 illustrates one exemplary implementation of a computing device that may be used in a system implementing techniques described herein, in accordance with some embodiments.

FIGS. 5A-5D are schematics showing the production of a preferred substrate for immobilizing leukocytes during imaging, according to an aspect of the invention; and FIG. 6 is a schematic showing a partial cross-section of a leukocyte bound to the substrate of FIG 5D.

DETAILED DESCRIPTION

For the assistance of the skilled reader, section I of this disclosure provides a general description of methods for imaging and characterizing single cells, before moving on in section II to give specific guidance about the imaging of leukocytes - in particular, T cells and CAR-T cells.

I. Introduction to Single-Cell Characterization and Cell Imaging Techniques

The present disclosure provides techniques for improving single-cell characterization methods. Some techniques described herein provide for the determining the molecular characteristics of cells in terms of both the number and spatial organization of biomarkers therein. In some aspects of the techniques described herein, diffraction limited super-resolution imaging is used to determine the molecular characteristics of cells, such as presence, abundance, and spatial distribution of biomarkers on cells of interest. In some aspects of the techniques described herein, cells may be classified according to higher-dimensional analysis of the molecular characteristics of individual cells. As set out in the background section above, single-cell characterization methods are of substantial value to many areas of biological and biomedical research. The parameters reported by such methods vary, but typically relate to the abundance of specific molecules in or on each cell (including proteins, nucleic acids, metabolic intermediates etc.) or morphological features of the cell or internal organelles (including cell size, granularity, dendricity etc.).

Single-cell characterization methods may provide single cell resolution of gene expression and biomarker density by identifying the organization of proteins within the cell and/or proteins associated with the cell membrane. The gene expression and biomarker density may be used to identify cellular activity, health, and/or status of the cell. Single cell characterization methods may quantify features of gene expression and/or biomarker density, such as quantifying the abundance of specific molecules in or on each cell. For example, single cell characterization methods may quantify proteins, nucleic acids, and/or metabolic intermediates. Additionally, or alternatively, single cell characterization methods may quantify morphological features of the cell and/or internal organelles. For example, single cell characterization methods may quantify cell size, granularity, and/or dendricity.

Relative to other single-cell characterization methods, flow cytometry is a high throughput method to acquire data on both biomarker abundance and cell morphology. In flow cytometry, hundreds of thousands of cells may be sampled by flowing cells through detection channels on the instrument. However, in some applications flow cytometry is limited by low sensitivity and low precision with regard to the characteristics of individual cells. Typically, to detect a specific biomarker by flow cytometry, thousands of copies of a molecule may be required because each cell is only measured for a very short period of time (e.g., on the millisecond scale, depending on the flow rate). Thus, the above-background signal is dependent on the integrated fluorescent intensity of a large number of molecules. Due to the integration of signal over a large number of molecules, the precision for detecting variations in the fluorescent emission form individual cells is low. The low precision is further exacerbated by the highly variable number of photons emitted by a single fluorophore within the detection time. As a result, flow cytometry may be limited to detecting large changes in biomarker numbers and even a change in factor of 10, 100, or 1000 copies/cell may not be reliable reported. However, biologically significant changes may occur at levels below the flow cytometry detection threshold. For example, low levels of gene expression or moderate changes in biomarker density can indicated differences in cellular status. Thus, the inventors have recognized and appreciated that more sensitive and quantitative approaches than flow cytometer may provide for earlier detection of biologically relevant changes for some applications.

Relative to flow cytometry, bulk analyses may provide increased sensitivity for a more quantitative approach to detecting biomarkers. Fluorescent microscopy methods have the potential to be significantly more sensitive to the presence of cellular biomarkers due to the fact that they collect far more light from each fluorophore and can include methods to overcome the stochastic differences in number of photons emitted by each molecule. Many microscopy methods that use high-NA objectives to collect light are sensitive to the level of single-molecules, meaning that in principle every molecule in a sample can be detected and quantified. For example, direct stochastic optical reconstruction microscopy (dSTORM) may be used to report the presence of CD 19 on the surface of myeloma cells with 3 -log greater sensitivity compared to flow cytometry. Such approaches offer substantial potential to report cell expression profiles at levels that are below the sensitivity of current single cell approaches but are nonetheless biologically meaningful.

However, bulk analyses do not provide results at the single-cell level (e.g., ELISA, quantitative proteomics, etc.). By contrast, single-cell transcriptomics (e.g., scRNAseq) offer some potential to provide results at a single cell level, however single-cell transcriptomics are limited by their shallow read depth (i.e., low sensitivity to gene expression with low concentrations of mRNA), relative to bulk transcriptomic analyses and the lack of absolute correction between transcript copy numbers and protein copy numbers mean that interpretation is highly limited.

The inventors have recognized and appreciated that the native configuration of proteins either within the cell or on the surface of the cell may be unique identifier of the type of cell and/or cellular status, like a fingerprint of the cell. However, beyond simple molecular counting, an additional parameter that is not presently accessible through conventional techniques is the nanoscale organization of molecules within or on the surface of cells. Diffraction-limited imaging may be used to detect the gross organization of molecules with a resolution of hundreds of nanometers, which may address questions relating to processes at the microscale - e.g., broadly where a protein is within the cell (nuclear, cytosolic, plasma membrane etc.). Furthermore, there are many high-throughput, low-resolution imaging methods that can be used to obtain such images, including variants of imaging flow cytometry. However, no such method provides information on the nanoscale organization of molecules in the native configuration in combination with single-molecule counting and process that information in a systematic, tractable manner. Nonetheless, the inventors have recognized and appreciated that there is substantial value in reporting molecular organization at the nanoscale, since this is the functionally relevant scale for most molecular biological processes. As such, the nanoscale organization of cellular molecules can be a key indicator of, for example, their activity, interactions, and regulation; and so also has value as a reporter of cellular activity, health, status etc.

The inventors have further recognized that processes which do not take care to preserve the native distribution of the nanoscale organization may cause reorganizations of molecules on the surface of or within the target cell. While some applications may not be disrupted by deviations from the native nanoscale organization of a cell, the inventors have recognized and appreciated that for some applications, capturing the native nanoscale organization of a cell may provide advantages in cell classification.

Therefore, the inventors have developed techniques for determining the nanoscale organization of molecules on the surface of and/or within a cell. Additionally, the inventors have developed processes for determining the nanoscale organization of molecules for a cell that preserves the native nanoscale organization. In some embodiments, the process to determine the nanoscale organization of molecules for a cell comprises applying a fixation agent to the cell while the cell is in a native configuration to produce a fixed cell; capturing the fixed cell on a surface of a sample substrate; and acquiring an imaging sequence of single molecule emission from the fixed cell.

As discussed above, the inventors have recognized that nanoscale organization of a cell may be used to determine the type of cell and/or to determine statuses or conditions of the cell. FIG. 1 A illustrates an exemplary collection of cells at the microscope scale. At the microscopic scale, as would typically be observed by an optical microscope, individual cells may be visible within a field of view. For example, cell 100 and cell 102 may be independently resolved using a bright- field imaging technique. However, while it is generally possible to resolve cells in an optical microscope, the nanoscopic features of a particular cell is not usually observable using bright- field imaging. FIG. IB illustrates an exemplary native distribution of surface proteins at the nanoscopic scale of cell 100. As shown in FIG. IB, a first surface protein 120, 122, and 123; a second surface protein 110, 112, 114, and 116; and a third surface protein 130 and 132 may be present on the surface of the cell. In some embodiments, the quantity and/or the density of a particular surface protein may be used to provide classification information about the cell. For example, three instances of the first surface protein and four instances of the second surface protein may be detected with a relatively low density on cell 100. However, two instances of the third surface protein may be detected at a relatively high density on cell 100. A cell classification based on the presence and density of the surface proteins may be used to classify attributes of cell 100.

FIG. 2A illustrates a method for acquiring the spatial organization data of cell structures, in accordance with some embodiments. Process 200 starts at block 202 by capturing a leukocyte on a sample substrate. Cell capture involves binding target cells to a substrate surface. Binding between the target cells and the substrate surface is facilitated by adhesive forces between the surface of the cells and the surface of the substrate. When adhesion between the cells and the surface is strong enough, the adhesive forces restrict cell movement immobilizing the cells on the substrate. In some embodiments, the substrate surface is an imaging-compatible surface. For example, a glass coverslip may be used as an imaging-compatible surface. In some embodiments, a binding layer may be disposed on the surface of the substrate to facilitate adhesion.

For some cell types (in particular leukocytes), the cell membrane may have a complex morphology, such as ruffling which creates a complex three-dimensional structure. When imaging cells which include such complex morphologies, it can be challenging to determine the density of a biomarker of interest because it is difficult to determine the surface area of the membrane based on a two-dimensional area observable in an image. As an example, nonadherence cell types, such as lymphocytes, have extremely ruffled morphology and are capable of forming complex membrane structures such as microvilli and pseudopodia. The inventors have recognized and appreciated that by utilizing capture methods which provide strong binding of cells to the imaging surface the three-dimensional morphology of the cells can be disrupted. The forces between the surface and the cell membrane may induce the cells to spread and form a large, flat contact area with the surface. From the contact area between the cell and the surface, the surface area of the cell membrane can be determined from the two-dimensional area observable in an image, enabling determinations of the density of biomarkers within the cell membrane from the two-dimensional image. Typical cell capture methodologies include a non-specific binding layer to capture cells. Binding layers may include a polymer layer adsorbed to the substrate. For example, positively charged polymers may be used to form the polymer layer when the target cells predominately include negatively charged headgroups of the cell plasma membrane. For example, poly-L-lysine (PLL), poly-D-lysine (PDL), chitosan, or nitrocellulose may be used as polymer layers for binding.

In other typical approaches, the substrate may be used without an additional binding layer for non-specific binding. Adhesions between the substrate surface and the target cell can be increased by modifying the substrate surface such that it is has a complimentary charge to that of the target cell membrane, e.g. reactions such as hydroxylation may be used to charge a glass substrate surface. For example, piranha solution, strong oxidizing agents, or plasma etching may be used to charge the substrate surface. As another example, the surface may be treated to add functional groups that are charged and/or polar, such as aldehydes.

In some embodiments, the cells may be removed from growth medium prior to capture. Some growth media, in which the target cells are grown, may contain serum and/or other complex solutions which may compete with the target cells to bind to the substrate surface. To avoid competition between the growth solution and the target cells during capture, the target cells may be removed from the growth solution through centrifugation and resuspension in a buffer solution which will not compete for surface binding to the substrate. For example, the target cells may be resuspended in phosphate buffer solution (PBS), HEPES buffer solution (HBS), or a solution of PIPES, HEPES, EGTA, and magnesium sulfate (PHEM).

In embodiments, capture includes a biomarker-specific binding layer to capture cells. The binding layer may include specific capture molecules, including antibodies, antibody derivatives, and/or aptamers to cause selective binding with target cells. For example, CD4 antibodies may be included in the binding layer to selectively bind CD4⁺ CAR-T cells. According to some aspects, it may be beneficial to separate cells prior to analysis. By including CD4 antibodies, the substrate can selectively capture CD4⁺ CAR-T cells from a solution of both CD4⁺ and CD8⁺ cells.

In embodiments, the selective binding surface is passivated to prevent non-specific attachment of cells beyond the target cell. For example, an uncharged passivation coating such as polyethylene glycol (PEG) or other non-specific adsorption of generic large molecules, such as bovine serum albumin (BSA).

In one embodiment, specific capture may be achieved by cleaning and silanizating an imagingcompatible glass surface. For example, the surface may be cleaned using a piranha solution, plasma cleaning, or a high pH concentrated salt solution treatment - such as potassium hydroxide. Once clean, the surface may be treated with a silanizing agent. For example, 2-aminopropyltriethosxysilane may be used to add aminosilane groups which covalently link to biotinylated N-hydroxysuccinimide (NHS)-PEG. The PEG-biotin passivates the surface to prevent non-specific deposition of contaminant particle. In some embodiments, the PEG-biotin may be used to facilitate coating of the glass with multivalent biotin-binding proteins. For example, streptavidin and/or neutravidin may be used to immobilize biotinylated cell-capture molecules.

Two factors should be considered when performing direct capture of cells onto an imaging surface. First, does the capture process influence the distribution of molecules on /in the cell. Second, does capture alter the status, such as the activation or health, of the cell. Molecules may be driven by size-exclusion out of the contact region between the cell and the imaging surface if their extracellular dimensions exceed those of the cell-capture surface separation. For example, molecules may be driven out of the cell contact region between the cell and the imaging surface by size-exclusion when their extracellular dimensions exceed those of the cell-capture surface separation. The size of the molecules and the nature of the capture method will impact the extent to which size-exclusion may result in changes to the distribution of molecules on and/or in the cell. For example, some immune cell types, T cells and basophils, may be activated by capture on PLL due to size-dependent exclusions of large, inhibitory phosphatases such as CD45 from the contact region. Capture can also lead to change in cell status through other means. For example, activation of surface receptors by capture antibodies or mechanical stress on the plasma membrane during stretching upon capture. In embodiments that seek to observe the native distribution of molecules on or within the cell, steps should be taken to ensure the distribution of molecules and status of the cell at the point of fixation are reliably preserved after cell capture to the surface. The capture method according to the invention used for capturing leukocytes is as set out in the summary of the invention, involving the use of a CD45 capture molecule.

Next, the process proceeds in block 204 by applying a fixation agent to a cell in its native configuration. A fixation agent preserves the cells from undergoing reorganization. As a result, a neutral fixation agent will preserve the configuration of a cell when the fixation agent is applied. To preserve the native configuration of the cell such that the native distribution of proteins can be observed.

Fixation can be performed with any suitable fixation agent which preserves the molecular distribution long enough to acquire images from which the distribution of a target molecule may be determined. A range of standard fixative agents may be used, such as formaldehyde, glutaraldehyde, or glyoxal. For example, a 4% paraformaldehyde in phosphate buffered saline (PBS) may be applied to the cells for ten minutes at room temperature to produce a robust fixation of the cells, without over-fixating and damaging the morphology and/or epitopes required for antibody recognition (if performed after fixation).

In some embodiments, a quenching solution may be applied to the cells following application of the fixation agent, when the fixation agent which is used generates autofluorescence. For example, a quenching solution of 1 mM glycine or sodium borohydride may be applied to the cells after fixation to quench autofluorescence which may produced by the quenching agent.

In some embodiments, the captured cell may be compressed to flatten out the cell membrane between a substrate and coverside for the cell. In some embodiments, the strength of the capture process may cause the cell to flatten across the substrate. In some embodiments, a combination of the capture strength and pressure may be used to facilitate flattening of the cell across the substrate. The inventors have recognized and appreciated that flattening of the cell membrane prior to imaging may enable the calculation of a target molecule’s density across the surface area of the cell. Without the flattening of the cell, the folds and wrinkles of the cell membrane may prevent accurate determinations of a density for the target molecule as a function of cell surface area. Next at block 206, an image sequence of the fixed cell is acquired, in accordance with some embodiments. Cell imaging may be performed by any suitable microscope which is configured to produce super-resolution images of a cell.

According to some aspects of the present technology, cell staining/labelling may be used when the target molecule is not fluorescent. Biomarker-specific probes may be used for cell staining. Staining may be performed on biomarkers of interest for quantification and/or identification. In some embodiments directed to quantification, the abundance of the biomarker may be of interest.

In some embodiments directed to identification, the presence of a particular biomarker may be of interest. For example, different biomarkers may indicate whether a T cell is a CD4 or CD8 T cell.

In some embodiments, both the presence and quantity of biomarkers may be of interest. For example, antibodies, aptamers, purified or synthetic ligands which are pre-conjugated to reporter molecules that are compatible with the acquisition techniques described herein may be used as biomarker-specific probes.

Staining/labelling may be performed either before or after cell capture. Pre-capture and postcapture staining may be preferable to ensure all biomarkers are equally stained, as probes may be restricted in their access to biomarkers in the cell-capture surface interface due to the close proximity of the surface and the cell plasma membrane. This will be dependent on the nature and size of the probe in use, and the strength of the capture method.

In some embodiments, when pre-capture staining/labelling is used, unbound probes may be removed prior to capture in order to prevent adsorption to the surface independently of the cell, which can contribute to background or noise. Non-specific capture methods may exacerbate the adhesion of unbound probes to the surface, relative to biomarker-specific methods. In some embodiments, unbound probes need to be removed, this can be achieved by straightforward rounds of washing through centrifugation or other cell washing approaches clear to those skilled in the art.

In some embodiments, when post-capture labelling is used, the passivated capture surface prevents non-specific deposition of conjugated probes. This may be achieved as part of the capture approach or after capture and fixation by blocking with BSA casein, serum, or other such blocking agents. For example, PEGylation during the capture approach may passivate the surface to prevent non-specific deposition of conjugated probes. As another example, using a blocking agent such as BSA, casein, serum or other such blocking agent may be used during fixation to passivate the surface to prevent non-specific deposition of conjugated probes.

Staining of intracellular biomarkers is facilitated by prior permeabilization of the cell, in which case fixation must be performed first in order to stabilize the non-lipid components of the cell. Many types of intracellular molecules may be of interest in the characterization of cells. For example, markers of activation, such as phopho-ZAP70 in T cells; signaling intermediates, such as diacyl glycerol; markers of metabolic activity, such as mitochondrial components; cell cycle status, such as cyclin-CDK complexes; cell health, such as caspases; or cell responses, such as intracellular cytokines. In some embodiments, a gentle semi-reversible detergent maybe used. For example, a 0.1% saponin solution. In some embodiments, a strong detergent can be used when more substantial membrane solubilization is required to reach target molecules within the cell. For example, a 0.3% solution of Triton-X 100 could be used.

In embodiments where molecules of interest are intrinsically labelled with appropriate molecules for detection such as transgenic fusions of target proteins with fluorescent proteins, orthogonal staining may not be required.

In some embodiments, the ONI Nanoimager® platform or other appropriate microscope may be used to acquire an image sequence of the fixed cell. There are several techniques which may be used to drive high-sensitivity quantification of biomarkers on samples prepared, as discussed herein. In some embodiments, acquisition can be automated to run directly following sample preparation, where finding focus and field of view, then subsequent data acquisition is run by the software without operator input. This is facilitated by the fact that the capture surface is entirely homogeneous so there is no need to identify specific features for examination. Accordingly, a simple autofocus, wherein the optimal z-position for a given TIRF angle is automatically determined, is sufficient to allow automated acquisition.

In some embodiments, this allows for automatic collection of multiple datasets, within one sample or across samples, which is then compatible with more high-throughput experiments. In some embodiments, this may be combined with automated sample collection methods. For example, periodic sampling of cell preparations within a bioreactor. The combined data may be analyzed in accordance with the methods described herein.

In embodiments directed to cell-surface molecules, this strategy has the added advantage that the focal plane of interest will be the same for all cells in a field of view, so separated acquisitions at different cell-specific focal planes are not required.

In some embodiments, single-molecule localization microscopy (SMLM) techniques may be used to detect emission from the target molecules. For example, direct stochastic reconstruction microscopy (dSTORM), point accumulation for imaging in nanoscale topology (PAINT), photoactiviation localization microscopy (PALM) and/or widefield/confocal microscopy may each be used.

For example, dSTORM techniques may be used to resolve individual molecules with a resolution of 10-20 nm, in accordance with some embodiments. dSTORM is a SMLM that is based on the temporal segregation of signals from individual fluorophores by the promotion of stochastic blinking of fluorophores such that only a subset is fluorescent at any one time. To achieve this, staining (as described herein) is undertaken with agents conjugated to dSTORM-compatible fluorophores. For example, AlexaFluor647, AlexaFluor568, Atto488 or another suitable fluorophore may be used. Acquisition may be performed in a dSTORM buffer, such as a a highly reducing, oxygen-scavenging solution. For example, ONI’s BCubed buffer. Image collection is performed using appropriate acquisition parameters for the approach used. For dSTORM, a typical acquisition cycle is 2000 frames at 33 ms/frame with sequential cycles for each channel. dSTORM can be used to report the presence of molecules in an ultrasensitive manner, as described herein.

Photoactivation localization microscopy (PALM) is an SMLM approach similar to dSTORM. Photoactivatable fluorophores (including organic e.g. PA-JF646, abberior CAGE 552 etc., or protein-based e.g. PA-GFP, Eos etc.) are used and spatially separated emission from individual molecules is achieved by low-level irradiation with the activation wavelength (typically in the UV range). This can be done with continuous activation at a low level, which leads to stochastic activation similar to the blinking behavior achieved in dSTORM, or through cyclical rounds of activation and detection in which case a subset of fluorophores is activated within each cycle. Localization of individual emission events is performed in an equivalent manner to dSTORM. As another example, PAINT techniques are also SMLM techniques that is based on detecting the blinking behavior produced through the transient binding and detachment of fluorophore- conjugated molecules to biomarkers or biomarker-targeted probes (e.g. antibodies, aptamers etc.) within the sample. For example, DNA-PAINT or peptide-PAINT may be used depending on the target molecule. In this manner, PAINT can be used in a similar way to dSTORM to facilitate ultrasensitive biomarker detection. In order to facilitate even more directly quantitative biomarker detection, PAINT events can be benchmarked against known standards with the assumption that probe binding kinetics are equivalent, thereby allowing absolute numbers of binding sites in the sample to be derived (so-called qPAINT). In cases where PAINT methods are to be used, imaging should be performed in a suitable PAINT -compatible buffer rather than a dSTORM buffer.

As yet another example, when using biomarkers to identify specific subpopulations of cells or cell status standard widefield or confocal fluorescence microscopy may be used. For example, to identify the difference between stains which may indicate the presence of CD4 vs CD8 on T cells, or to identify cell status using markers of cell viability. In such cases, image acquisition is used to determine the presence or absence of the given marker, rather than providing high-sensitivity detection and quantification. For purposes of determining the presence or absence or a marker, standard widefield or confocal fluorescence microscopy is sufficient. This can be easily performed in conjugation with the more quantitative approaches discussed above by simply altering the acquisition mode during data collection. Biomarkers being reported in this manner do not need to be imaged at the basal surface of the cell membrane, and so imaging can be performed at a more equatorial focal plane in order to minimize any effects of photobleaching caused by imaging other biomarkers at the basal plane. In some cases, imaging at a higher focal plane may be inherently required, such as in cases where markers are within specific organelles (e.g. viability markers targeting nuclear DNA). Biomarkers used for this kind of reporting can still be imaged using any of the approaches described above if preferable - e.g. by DNA-PAINT in order to permit sequential probing and increase the breadth of effective channels available.

Acquisition methods may be used to detect multiple molecules of interest in parallel, in accordance with some embodiments. Detecting multiple molecules of interest in parallel may provide for more granular identification of different cell populations, deeper analysis of proteomic changes in response to stimulus relative to stimulus, when compared to detecting a single-molecule of interest. To facilitate detection of multiple molecules in parallel, the signals from different molecules of interest may be multiplexed together. Multiplexing may involve collecting signals with different spectral and/or temporal emission profiles. Detection of multiplexed signals may be enabled by spectroscopic detection. Spectroscopic detection may be achieved through by ratiometric comparison of fluorophore emission above and below a specific optical cut off wavelength or through characterization of the full emission spectrum. The full emission spectrum may be characterized through the use of prism or grating based spectral separation or using a series of discrete bandpass filters.

Multiple molecules of interest may be detected sequentially, in accordance with some embodiments. To facilitate sequential detection of multiple molecules of interest, sequential rounds of staining/labelling and imaging using different probes may be used for imaging. The sample may be cleared of a given fluorophore prior to re-staining with a new probe bearing the same fluorophore. For example, an exchange PAINT technique may be used. In exchange PAINT, the imager oligonucleotide used in DNA-PAINT are sequentially removed and replaced with strands of different sequence specificity.

In some embodiments, sequential and/or parallel detection techniques for detecting multiple molecules of interest retains single-molecule detection sensitivity. For example, dSTORM, PALM and PAINT techniques are compatible with sequential and parallel detection techniques. Similarly, widefield or confocal fluorescence microscopy is compatible with sequential and parallel detection techniques.

Next at block 208, the spatial organization of cell structures is determined, in accordance with some embodiments. To determine the spatial organization of cell structures, a single-molecule localization algorithm is used to identify localization events with the acquired images. In embodiments that include sequential and/or parallel detection techniques, localizations are linked to their corresponding identity according to their relevant parameter. For example, when using spectrally distinct fluorophores localizations may be linked to a particular emission wavelength. As another example, when sequential detecting multiple molecules of interest, localizations may be linked to a particular frame number associated with the fluorophore used in imaging for that frame.

According to some aspects of the technology described herein, fluorescence signals may be acquired using total internal reflection fluorescence microscopy (TIRF). In TIRF, a collection lens or microscope objective is configured to receive fluorescence light from the sample and transmit the received light through an optical path to a detection device.

A super-resolution fluorescence microscopy system suitable for carrying out TIRF microscopy is shown in Figure 2D. Figure 2D shows a test specimen 251 mounted on motorized stage 252. The test specimen consists of a sample of fixed leukocytes prepared according to the methods of the invention, immersed in an imaging buffer. The imaging buffer is compatible with dSTORM, containing a reducing agent (e.g. a primary thiol such as P-mercaptoethanol (BME), mercaptoethylamine (MEA), dithiothreitol (DTT) or L-glutathione) and an oxygen scavenging system (e.g. the combination of glucose oxidase and catalase, or the combination of protocatechuic acid (PCA) and protocatechuic dioxygenase (PCD)). The leukocytes have been labelled with a dSTORM compatible fluorescent probe having specificity to a biomarker on the cell surface, and have been fixed prior to imaging. The dSTORM compatible fluorescent probe includes a photoswitchable fluorophore, which is able to switch from a dark state to an emissive state.

The sample 251 is interrogated by Total Internal Reflection Fluorescence Microscopy (TIRFM) system 253. In the TIRFM system 253, excitation beam 254 from laser 255 is reflected by dichroic mirror 256 so as to pass through the edge of objective lens 257, and totally internally reflect off the top surface of the coverslip on which the sample is placed. This creates an evanescent field, which switches a small proportion of the photoswitchable fluorescent probes from a dark to an emissive state. Fluorescence emission from the emissive fluorescent probes is then collected by objective lens 257 and passes through dichroic mirror 256 and optical filter 258 before being detected on EMCCD camera 259. Signal from the emissive fluorescent probes then disappears, either due to the fluorophore switching back to a dark state or photobleaching. Through control of conditions (in particular laser power), the density of photoactivated fluorescent markers in each image recorded by the camera is such as to allow individual fluorescent markers to be identified as separate points. By acquiring multiple images, it is possible to gradually construct an image of individual fluorescent markers across the cell surface.

An alternative setup for TIRF imaging is shown in FIG. 2B with microscope objective 216 configured to collection fluorescence emission 214 from cell 210. Here, the excitation light is incident on the sample from the side of the sample which does not face the collection lens or microscope objective. Furthermore, the excitation light is incident at an angle as to cause total internal reflection. In this way, evanescent waves propagating from the reflection of the excitation light will have a penetration depth into a sample disposed on the surface that the excitation light is reflecting from. In this way the excitation light is reflected away from the collection lens or microscope objective and, by extension, the detection optics.

FIG. 2C illustrates the excitation geometry for the microscope configuration illustrated in FIG. 2B. As shown in FIG. 2C, excitation light 220 is incident on cell 210 from the bottom surface, opposite microscope objective 216 that is located above cell 210. At the surface of the substrate upon which cell 210 is disposed, excitation light 220 undergoes total internal reflection, resulting in reflected excitation light 222 and evanescent radiation which will propagate a penetration depth 224 into cell 210. As a result of the excitation from the evanescent radiation, fluorophores in cell 210 can become energetically excited, causing fluorescent emission 214. Although illustrated as a single reflection at the interface in FIG. 2C, many internal reflections may occur between the two surfaces of the substrate.

FIG. 3 illustrates process 300 for the initial processing and analysis of single-molecule localization data, in accordance with some embodiments. Prior to process 300, single-molecule localization microscopy data may be acquired in accordance with process 200, described above for acquiring the spatial organization data of cell structures.

Process 300 begins at block 302 by determining location events. Location events are signals which may be indicative of the location of a molecule on the surface of or within the target cell. The location events may be determined from an unprocessed image and/or series of images, such as an image stack. In some embodiments, identification of localization events may be determined using the ONI NimOS single molecule localization algorithm or other localization tool configured to identify positions within an image which captured fluorescence signals. In embodiments where multiplexing is used, localizations should be linked to their corresponding identity according to the relevant parameters, such as spectral identity and/or frame index.

At block 304, when multiple images are acquired a drift correction may be performed, in accordance with some embodiments. For example, in connection with different staining cycles, different times following photo-excitation of fluorophores, and/or for purposes of repeating measurements performing a drift correction may register features from across multiple images together such that they share a common coordinate system for determining the position of signals captured by each image.

At block 306, data may be filtered by quality metrics, in accordance with some embodiments. Quality metrics which describe the quality and/or confidence of determinations based on the captured data may be determined for each image and each signal captured within each respective image. After determining quality metrics for captured data, the data may be filtered. For example, acquired images and/or spectra may be filtered by number of photons detected, precision of localization, a sigma parameter resulting from the fit, and/or a goodness of fit metric.

At block 308, localizations which are determined from different frames may be averaged together as a temporal grouping, in accordance with some embodiments. In a temporal grouping, the properties of the representative grouping are the averaged values of the properties of the individual localizations from representative frames. For example, each localization which has been identified within a grouping region in different frames may be grouped together and the average position and/or photon count may be used for the property of that representative grouping. In some embodiments, the grouping region may be between 20 nm to 400 nm in diameter. In some embodiments, the grouping region may bet between 40 nm to 200 nm in diameter.

In some embodiments, gap frames will be present where no fluorophore is observed. For example, there may be 0 to 3 gap frames between frames where the fluorophore is detected. When zero gap frames are permitted, fluorophores which are spatially detectable near each other in separate frames will be considered separate fluorophores if a gap frame is detected between frames where emission is observed. The number of permitted gap frames may be determined by a wait time divided by an exposure time. For example, for a wait time of 100 ms and an exposure time of 33 ms, three gap frames may be permitted.

At block 310, after temporal grouping, the temporal groupings may be filtered to be excluded from the data set according to the groupings which appear in more than a chosen number of frames. Fluorophores may be characterized by a particular switch off time or fluorescent lifetime. When a detected signal exceeds the switch off time or fluorescent lifetime may be excluded as an artefact. In some embodiments, a grouped localization which appears in more than a chosen number of frames will be excluded. For example, a grouped localization which appears in more than 20 frames may be excluded from further analysis. As another example, a grouped localization which appears in more than 10 frames may be excluded from further analysis.

At block 312, a density of molecules may be determined on a cell-by-cell basis. Prior to determining a density of molecules, individual cells may be identified within the image field of view. Accordingly, regions of the image corresponding to each cell are identified and are then treated as discrete entities for downstream analysis. In some embodiments, individual cells are differentiated from empty regions of the image and from other cells using a feature-detection algorithm. For example, a machine learning-based identification of specific cellular features may be used to identify individual cells based on a bright field microscopy image of the cell. As another example, an intensity threshold may be used to identify regions of an image which are associated with a cell. The intensity threshold may be associated with a particular color channel or may be associated with an overall intensity threshold, as aspects of the technology described herein are not limited in this respect.

In some embodiments, cell identification techniques may be combined with interferometric imaging techniques. For example, interference reflection microscopy may be used to determine the region of contact between a cell and the imaging surface by identifying interference changes across the interferometric image.

To determine the density of molecules for a cell, the number of molecules emitting fluorescence is counted. In some embodiments, single-molecule localization microscopy data is used to derive the number of target molecules from the temporally grouped localizations. The inventors have recognized and appreciated that basing the number of target molecules on the temporally grouped localizations provides a crude estimate of protein numbers but may be prone to overestimation due to the repeated sampling of individual fluorophores. However, crude estimation of protein numbers may be used to identify the presence or absence of a target molecules because each emitting molecule is captured in the temporally grouped localization.

In some embodiments, the number of fluorophores within a given area can be derived from information associated with the acquisition method. For example, qPAINT determines molecular counts from the standard kinetics of probe binding and dissociation, as discussed herein. As another example, dSTORM uses physical properties of fluorophores, such as photo switching kinetics and duty cycle, can be used to determine molecular counts. At block 314 the spatial information is extracted, in accordance with some embodiments. Spatial information that describes the spatial organization of molecules within the image may include clustering metrics determined by a clustering algorithm. In some embodiments, a bespoke clustering algorithm may be used to determine clustering metrics representative of the organization of molecules within the image. Based on the clustering metrics, each molecule may be identified as associated with a stable cluster of localizations or as non-clustered. The clustering metrics may provide quantified values for several cluster-related parameters including cluster dimensions/shape (circularity, length, skew, boundary length etc.); cluster area; molecule density within a cluster; and/or number of molecules/cluster. Other clustering parameters may be included as aspects of the technology described herein are not limited in this respect.

In some embodiments, several metrics on a scale above individual clusters are also reported. For example, the fraction of clustered vs non-clustered molecules across the whole cell contact; the density of non-clustered molecules; the distance between clusters; the distance between clusters and non-clustered molecules; and/or the heterogeneity of these parameters across the whole cell contact, such as clusters of clusters or regions of low/high non-clustered molecule density.

In some embodiments, multiple biomarkers imaged may be in the same cell. When cells include multiple biomarkers, the spatial information may include cross-channel comparisons. For example, cross-channel comparisons may include the degree of cross-channel cluster overlap; the distance between molecules/clusters across channels; and/or the correlation of density of multiple molecules within the same cluster group.

Following the end of process 300, the data derived from the initial processing and analysis is further analyzed using multidimensional analysis, in accordance with some embodiments. To facilitate visualization of data, the relationships between large numbers of dimensions are flattened into two or three dimensions for ease of interpretation. In some embodiments, a via nonlinear reduction process may be used. Non-linear reduction processes of imaging-derived metrics allows cells with similar characteristics to be clustered together and hence cell populations to be separated within a heterogeneous sample. For example, a t-distributed stochastic neighbor embedding (t-SNE) process may be used. In other embodiments, other non-linear processes, such as UMAP, or linear processes, such as PC A, approaches may be used, as aspects of the technology described herein is not limited in this respect. Image acquisition can be performed using any number of channels. However, the inventors have recognized and appreciated that for every additional channel, several additional dimensions are collected, which may be used in downstream analysis. In some embodiments, a single biomarker is imaged in one channel. In this case, the minimum number of dimensions extracted is still sufficient to benefit from multidimensional analysis. In some embodiments, this includes dimensions relating to overall cluster morphology: size, circularity, skew, and other cluster metrics as described herein.

Additionally, or alternatively, dimensions relating to biomarker abundance and spatial distribution: absolute number, nearest-neighbor distance, extent of surface coverage, clusteredness may be included. The combination of these dimensions (e.g., size and number to give overall density; circularity and clusteredness to give polarization etc.) adds further dimensions.

In embodiments that include multiple biomarkers, biomarker-specific channels are added for each additional biomarker. With each added biomarker, the dimensions relating to biomarker abundance and spatial distribution are added for each subsequent biomarker. However, with the addition of more biomarkers, relative dimensions between each biomarker may also be considered. Relative dimensions may also be referred to as second-order dimensions. For embodiments that include multiple biomarkers, relative dimensions such as nearest-neighbor distance between biomarkers A and B; colocalization between biomarkers A and B; and/or ratio of relative biomarker abundance. For example, a 3 -colour dSTORM experiment including 3 biomarker channels would generate a minimum of 21 basic measurable dimensions and many more second-order dimensions. Dimensionality can be further increased through the use of higher-order functions or complex spatial analyses, such as Ripley’s K or Pair Correlation Functions. In a higher-order or complex spatial analysis approach, the first-order function would describe overall biomarker density, the second-order function would describe clustering (e.g. Ripley’s K), and the third-order would describe bandedness or equivalent matrix-derived features.

When bright-field microscopy images are used to identify individual cells within an image, the bright-field images may be used to segregate populations within a heterogeneous sample. Segregating the populations for analysis within a heterogeneous sample may provide for more precise interrogation of key cells of interest. For example, within a heterogeneous T cell sample there will be CD4+ and CD8+ T cells and one or both of these markers can be stained and imaged in a diffraction-limited manner. Following analysis as described above, the additional parameter of intensity in these reporter channels can be assigned to each cell and thresholds then applied in order to gate cells into a specific population. In this case, the CD4+ cells may be identified by applying a minimum threshold for CD4 intensity and a maximum threshold for CD8 intensity, and vice versa for the CD8+ population. Increasingly complex gating strategies can be used as additional markers are added. This is comparable to the gating strategy commonly employed during the analysis of flow cytometry data.

The analysis may also be paired with downstream analysis, in accordance with some embodiments. Following the identification of different cell subpopulations within a heterogeneous sample using higher-dimensional analysis, the coordinate positions of each cell can be used in order to retrieve physical material from the population of interest for downstream analysis. This can be achieved in a number of ways, for example, individual cells can be destroyed using laser ablation and the material collected for downstream mass spectrometry (e.g. for metabolomic analysis). Alternatively, specific cells can be released from the capture surface for collection if UV-photocleavable linkers are incorporated between the capture protein (e.g. antibodies) and the capture surface, with the coordinate positions of only the cells of interest targeted with a focused UV light source. Similarly, UV-photocleavable linkers can be introduced between biomarker probes (e.g. antibodies) and multiplexed reporters, such as oligonucleotide barcodes for downstream sequencing. The isolation of material from the specific cell subpopulation of interest has a significant advantage over bulk whole-sample analysis, which does not allow interrogation of cell heterogeneity or the contributions of a particular subpopulation to the total signal.

II, CAR-T workflow

In addition to the description of process 200 above, which describes the process for acquiring the spatial organization data of cell structures, for applications that are focused on chimeric antigen receptor (CAR) T cells the inventors have recognized and appreciated that additional factors may be associated with acquiring the spatial organization data in addition to or as an alternative to the factors described above with reference to FIG. 2A and process 200.

Capture The method of CAR-T cell capture will determine the status of the cell at the point of imaging, which may have a significant impact on the data generated. For example, capture methods that lead to increased CAR phosphorylation may increase CAR clustering or reduce CAR surface density due to internalization. Therefore, capture can be performed with the aim of either preserving the existing status of the cell, or of imposing a specific, controlled new status.

To preserve existing cell status, the capture method should preserve the distribution of molecules at the cell surface or interfere with the kinase-phosphatase balance that determines the extent of membrane protein phosphorylation to produce a desired status, in accordance with some embodiments. To facilitate control over the distribution of molecules at the cell surface, a minimum separation distance must be maintained between the basal cell membrane and the imaging surface. This is because the phosphorylation status of many lymphoid surface receptors (including TCR, CAR, and costimulatory/coinhibitory receptors) is sensitive to the passive segregation of large molecules from the contact region. Many such large molecules (e.g., CD45, CD148) are highly active tyrosine phosphatases that both dephosphorylate signaling motifs in surface receptors, but also contribute to the regulation of kinases (e.g., Src kinases) that phosphorylate the same motifs. If large molecules are excluded from the contact region then surface receptors within this region will spontaneously become more heavily phosphorylated. For example, if the cell-imaging surface distance is the approximate size of the extracellular domains, approximately >20-3 Onm, the large molecules may be excluded from the contact region promoting heavy phosphorylation.

In embodiments, capture reagents that target large molecules at the cell surface to hold the cell in place without bringing it too close to the surface may be used to maintain the captured T cell surface sufficiently far from the imaging surface, avoiding heavy phosphorylation. For example, one such molecule is CD45, which is ubiquitously and highly expressed on all T cells (and indeed all leukocytes - so this method is applicable to all immune cells) and has been used successfully to capture T cells without activation. Anti-CD45 (or capture molecules targeting other large surface antigens) can be immobilized on the capture surface directly through chemical adsorption or indirectly through the initial deposition of a secondary capture molecule. This could be a secondary antibody that targets the anti-CD45 in a species-specific manner, or an affinity probe against a specific affinity tag (e.g. neutravidin/streptavidin capturing biotinylated anti-CD45). In some embodiments, these approaches may be in combination with each other. For example, neutravidin on the imaging surface may be used to capture a biotinylated secondary antibody, which is used to capture anti-CD45. The addition of molecules between the anti-CD45 and the imaging surface serves to add further distance between the cell surface and the imaging surface, which may prevent capture-dependent activation.

Additionally, or alternatively, as set out above, a passivation reagent may be used between the imaging surface and the cell. For example, the imaging surface can be coated with PEG linked to biotin, which is then coated with neutravidin (through the biotin moiety), which in turn captures biotinylated anti-CD45. Whilst the PEG layer represents a surface from which the cell is kept removed, it is a much less rigid and less dense layer compared to most imaging surfaces (typically glass) and so is less likely to promote molecular segregation at the captured cell surface.

In some embodiments, additional considerations for anti-CD45 (or equivalent target) based capture may be included. The compatibility between the capture antibody and isotypes of the target that are present on all cells of interest may be included. For example, CD45 has several isotypes and not all structural elements are conserved between all isotypes, so the antibody should target common elements or elements only of specific isotypes of interest. Similarly, the binding epitope on the target molecule should be located distally to the cell membrane in order to maximize the separation distance between cell and imaging surface. The geometry of binding must be compatible with efficient capture (ideally the antibody and target should bind end-on to one another) and in a manner that does not alter the native orientation of the target relative to the cell membrane.

The extent of capture (i.e., the number of CD45 or equivalent molecules on each cell that are engaged by the capture molecule) should be the minimal number required to allow efficient capture. Excessive engagement of CD45 will enrich it within the contact area and thereby cause a reduction in surface protein tyrosine phosphorylation due to increased local phosphatase activity. The density of capture molecules can be titrated by reducing the underlying density of secondary capture elements. For example, when passivating with PEG-biotin, non-biotinylated PEG can be included during the coating step to reduce the biotin concentration on the passivated surface. Additionally, or alternatively, competitor molecules can be included during the antibody-loading step to reduce capture antibody density (e.g., irrelevant antibodies of the same isotype if captured by secondary antibodies; or free biotin if captured through neutravidin). For example, a 1:40 PEG-biotin:PEG and then a 1 : 10 molar ratio of anti-CD45-biotin:free biotin may be used. The use of a passivated PEG surface here may also improves the signal to noise, following staining, as it may minimizes non-specific probe deposition on the imaging surface.

Figures 5A-5D and Figure 6 show a preferred implementation for immobilizing CAR-T cells according to the present invention. In Figure 5A a glass slide 501 has been cleaned using piranha solution, with functional groups 503 added. The functional groups 503 are provided through reaction of hydroxyl groups on the glass slide 501 with 3 -aminopropyltri ethoxysilane. In Figure 5B, the glass slide has been treated with a mixture of PEG 205 and biotinylated PEG 207, both of which have reacted with the functional groups 503 on the glass slide so as to become covalently bonded to the slide. Next, neutravidin 509 is added as shown in Figure 5C, before addition of biotinylated anti-CD45 antibody 511 in Figure 5D.

Figure 6 shows the slide of Figure 5D after addition of a CD4⁺ T cell to the surface. This shows that anti-CD45 antibody has bound to the distal end of CD45 molecules 603. The attachment through CD45 means that other molecules on the cell surface, such as TCR cluster 607 and CD4 molecule 605 can move through the contact region unimpeded.

In embodiments which seek to capture the cell in the native state, capture using anti-CD45, or equivalent molecules, should be undertaken in conditions as close to resting culture conditions as possible. The method is compatible with capture in complete growth medium (i.e., containing serum) as it is not blocked by non-specific interactions from proteins or other molecules in the solution.

In embodiments which seek to observe how a cell responds to specific conditions rather than determine the underlying status of the cell, the cell capture is configured to induce a specific cell status. For example, inducing an activated phenotype by triggering signaling through the CAR can give insights into its sensitivity, responsiveness, effector phenotype etc. This can broadly be achieved in three ways including: immobile phase, mobile phase, solution phase, or combinations thereof. Immobile phase methods involved the deposition of immobile elements on the imaging surface that engage molecules on the T cell surface and induce a response. In some embodiments, this may include affinity probes (e.g., antibodies), recombinant ligands (e.g., CD80 for CD28, PDL1 for PD1 etc.), and non-native ligands (e.g., superantigens), and can be targeted against a wide range of surface proteins including activatory receptors (e.g., CAR, TCR, CD28, 0X40, 4- 1BB etc.), inhibitory receptors (e.g., PD1, Tim3, LAG3, HVEM, CTLA4 etc.), adhesion molecules (e.g., LFA1, CD2 etc.), and/or coreceptors (e.g., CD4, CD8), antigen-presenting molecules (e.g., MHCI, MHCII, CD1). These molecules can be immobilized using the same methods as described herein for anti-CD45, and can be combined in any number of ways to generate the desired phenotype. It is possible to compare similar surfaces in order to elucidate specific phenotypes. For example, surface 1 engages CAR alone, surface 2 engages TCR alone, surface 3 engages CAR and CD28, surface 4 engages TCR and CD28, this would allow the determination of the relative impact of costimulation by CD28 on the potency of signaling through TCR and CAR.

In some embodiments, micropatteming (e.g., using microcontact printing, localized photouncaging etc.) to restrict different capture molecules to different regions of the capture surface. This may be used to provide more subtle insights into the mechanistic details of cell behavior - e.g., if a response differs when two capture molecules are colocalized vs segregated, then the response may indicate the spatial requirements of the underlying signaling mechanism (i.e., do the two targets undergo distance-dependent cross talk).

Mobile phase methods may be used to mimic the 2-dimensional diffusion of molecules that occur on the surface of cells that would interact with T cells in vivo. Mobile phase methods involve the incorporation of equivalent capture molecules (as the immobile phase methods) on a mobile or semi-mobile substrate. This is most easily achieved using a supported lipid bilayer as a fluid artificial membrane, onto which capture molecules can be loaded. Mobile phase methods may provide an improved reconstitution of the natural dynamics of the T cell activation process - e.g., allowing proteins to cluster without the restriction of being bound to an immobile capture molecule (see discussion of Immunological synapse features, below). Mobile phase methods can be combined with immobile phase methods by partitioning the cell surface into mobile and immobile regions - e.g., by loading spatial separated nanoparticles onto the capture surface as docking sites for a subset of capture molecules and then surrounding with supported bilayer as a substrate for mobile capture molecules.

In some embodiments, solution phase methods may be used. Solution phase methods involve the capture of the T cell using an immobile or mobile phase method (including using non-triggering approaches such as anti-CD45) and then adding to the solution soluble effector molecules to affect cell status. Such effectors include soluble ligands (e.g., pMHC, CD80, CAR ligand etc.), anti-receptor antibodies, cytokines, chemokines, toxins, cytotoxic effectors, small molecule modulators (e.g., enzyme inhibitors etc.), and chemical effectors (e.g., reducing agents, pH changes). Equivalent molecules can also be used in aggregated formats (e.g., tetramerized through streptavidin linkers) or on solid supports (e.g., polystyrene/silica beads).

III B Fixation

The method of fixation may impact the preservation of the native distribution of molecules across the surface. In some embodiments, generic cell fixation is undertaken with common fixatives in simple buffers - e.g., 4% PFA in PBS. To preserve the pre-fixation state of molecules in CAR-T cells, stabilization of the underlying cytoskeleton should occur rapidly in the first steps of fixation. Without rapid stabilization in the first steps of fixation, the significant interactions between the cytoskeleton and surface proteins (through direct and indirect interactions, and cytoskeleton-dependent membrane-partitioning) will disrupt the preservation of the preservation of the cytoskeleton. A poorly preserved cytoskeleton leads to poor retention of pre-fixation membrane protein distribution. For this reason, fixation conditions that preserve cytoskeletal integrity should be used. For example, this may be achieved by using cytoskeletal-preserving buffers such as PEM (PIPES, EGTA, MgCb) or PHEM (PIPES, HEPES, EGTA, MgCb) as the base buffer into which the fixative is diluted. Additionally, a solution of 4% PFA in PEM/PHEM may be a suitable buffer for this purpose. Additional fixative agents such as 0.1% glutaraldehyde can be added on top of this to more thoroughly fix the cell surface in the case of difficult-to-fix targets, such as GPI-anchored proteins.

III.C Staining

The method of staining may impact the observable fluorescent signals. Staining in this workflow is typically achieved after cell capture and fixation, so all probes used should be compatible with fixed targets. In some embodiments, staining of live cells before or during capture may be possible when the staining method is part of the capture/activation approach. For example, fluorescently conjugated proteins on supported lipid bilayers, or antibodies that both stain and activate a target may be used to stain as part of the capture/activation approaches. However, in embodiments configured to capture target live cells in their native, unperturbed state, stain targets should not be used as the method of staining is likely to influence their behavior.

Staining of surface markers in CAR-T cells can be achieved using standard affinity probe methods as with any other cell type, such as antibodies. In some embodiments, the main nonstandard target for staining is the CAR itself, for which affinity probes may not be available due to the clonal nature of the CAR. CARs can be probes in a range of ways including anti-idiotype antibodies, soluble ligands, and antibody-binding proteins (e.g., protein L). The inventors have recognized and appreciated that when using CARs as probes, methods should be compatible with the geometry of the cell-imaging surface relationship. For example, if the size of the CAR-probe complex is larger than the gap between the cell and the imaging surface then the stain will be less effective. Therefore, rather than using full-length recombinant CAR ligands, some embodiments may benefit from using truncated versions that include only the structural elements that contribute to CAR recognition. The inventors have further recognized and appreciated that this is increasingly important the larger the CAR ligand is. For example, the CD19 extracellular domain consists of 2 Ig domains - approx. 7nm length total, whereas CD22 consists of up to 7 Ig domains plus a carbohydrate recognition domain - approx. 25nm length total. The way in which CARs operate as a probe in vivo (i.e., binding to targets on opposing cell membranes) means that most will have a ligand binding geometry that points the C-terminus of the CAR away from the ligand, which exacerbates the above effect. For the same reason, smaller probes are preferable when staining any target at the basal surface of the cell (i.e., Fab, nanobody, affibody, aptamer etc. may be better than whole Abs or full ligands). If the target is intracellular, permeabilization may be required to increase the permeability of the cell membrane for the probes to enter the cell.

In some embodiments, sequence-specific nucleic acid staining is also possible with this approach. This can be achieved using in-situ hybridization of fluorescently-conjugated oligonucleotides to target sequences of interest in the cell. For example, this may include both DNA- and RNA- based targets, including mRNA, miRNA, tRNA, rRNA, snRNA, siRNA, snoRNA, piRNA, tsRNA, eRNAs, and srRNA etc.

III.D Acquisition

Image acquisition may be conducted in accordance with any of the imaging methods described herein. However, the relevant acquisition method may vary according to target and the information expected from the target (e.g., quantification vs cellular phenotyping).

IV. CAR-T-specific targets

The following non-limited examples of molecules/features represent relevant targets for imaging in the context of CAR-T biology. The analysis and interpretation of the molecules/features in the context of CAR-T biology may be determined using the methods and processes described herein. Examples of cell surface proteins include, CAR, TCRab, TCRgd, CD3, CD28, CD160, TIGIT, CD45, Lek, ICOS, CD27, HVEM, CD40-L, 4-1BB, 0X40, DR3, GITR, CD30, SLAM, CD2, CD226, CTLA-4, BTLA, LAIR1, CD244, PD-L1, TIM1, CD84, CRACC, CD27, LIGHT, TIM3, LAG3, LFA1, and PDl.

Examples of recruited effector molecules include, ZAP70, SHP1, SHP2, LAT, PKCtheta (+ PTMs of all of these - e.g., pTyr-ZAP70).

Examples of Nucleic acid targets include, telomeres, modified gene integration sites (e.g., CAR- encoding gene copy number), mRNA of modified genes (both transduced - e.g., CAR, eTCR etc.

- and disrupted - e.g., PD1 knockout).

Examples of secreted soluble molecules, include cytokines (e.g., IL2, IL4, IL5, IL6, IL7, IL10, IL13, IL15, IL17, IL21, IL22, IFNg, TNFa, TGFb etc.), cytotoxic effectors (e.g., perforin, granzyme, granulysin etc.), and engineered secreted molecules (e.g., immune checkpoint inhibitors etc.)

Examples of secreted particles include, extracellular vesicles (e.g., CD9, CD63, CD81), supramolecular attack particles (e.g., perforin, granzyme, thrombospondin- 1) - with any secreted species (both soluble and particles) it is possible to measure both their particle-level features (e.g., composition, size etc.) and cell-level metrics (e.g., number released per cell, heterogeneity per cell, site of release within the cell etc.), which may be informative for understanding the efficacy of their delivery.

Examples of morphological features include, cell-imaging surface contact area and symmetry, cell volume, cytoskeletal organization (e.g., actin, tubulin, myosin), cell polarization, mitochondrial distribution, mitochondrial fragmentation, nucleus volume, Golgi integrity (GM130, TGN64, ERGIC), membrane integrity/morphology (membrane dyes - e.g., Cell Mask), MTOC position, features of immunological synapse/kinase (for additional details, see immunological synapse features below).

Examples of intracellular features, include location of RNA species, protein location (i.e., retained in ER vs surface trafficked), degree of protein aggregation/misfolding, histone modifications (methylation, acetylation, ubiquitination), biomolecular condensate structure and morphology.

The following targets represent exemplary molecules that can be used as cellular phenotypic reporters (i.e., non-characterized) in CAR-T cells: T cell/CAR-T cell indicators, such as CD3, CAR.

Examples of cell lineage indicators include, CD4, CD8, TCRalpha, TCRbeta, TCRgamma, and TCRdelta (including specific idiotypes of the different TCR chains - e.g., Vgamma9-delta2).

Examples of effector population markers include, Thl markers (surface: e.g., CXCR3, CCR5, IL12Rb2, IL27Ra, IFNgR2, cytosolic: e.g., IFNg, nuclear: e.g., STAT1, STAT4, Tbet), Th2 markers (surface e.g., CCR4, CCR3, CCR8, IL1R4, cytosolic: e.g., IL4, nuclear: e.g. STAT5, STAT6, GATA3) Thl7 markers (surface: e.g., CCR6, CD161, cytosolic: e.g., IL17, IL26, nuclear: e.g., STAT3, RORgt), Treg markers (surface: e.g., CD25, CD127, CD152, CTLA4, GITR, 0X40, cytosolic: e.g., TGFb, and nuclear: e.g., FOXP3, STAT5).

Example of markers of naive vs activated and effector vs memory: e.g., CD45RA, CD27, CCR7, CD95, CXCR5, PD1, CD38, ICOS, IL2, CD45RO, CD62L, CD44, CD69, CD25, CD103, CD71, HLA-DR, CD62L, NFkB, NF AT, and CD95.

Examples of exhaustion/senescence markers include, PD1, TIM3, LAG3, CTLA4, CD57, KLRG1, and TOX.

Examples of cell viability markers include, Sytox, Annexin V, and propidium iodide.

Examples of cell cycle indicators include, Cyclin A, Cyclin B, Cyclin D, Cyclin E, and Ki67.

The exemplary markers/features described herein can be assessed in isolation or in combination, as aspects of the technology described herein is not limited in this respect. When measures in combination, feature metrics that relate to the interplay between the individual targets can be obtained - e.g., distance between copies of target A vs target B, relative numbers of multiple targets etc. The reporting of the absence of these markers/features is also informative. For example, in cases where cells have been engineered to lose expression of a given protein (e.g., PD1 disruption by CRISPR-Cas9) then the correlation of phenotypes with the desired absence of a target will also be informative.

V, Immunological synapse features

The immunological synapse is the structure formed by T cells and other lymphocytes upon engagement of a target cell. The immunological synapse is characterized by the tightly regulated organization of molecules into spatially distinct compartments of the cell surface. Assessing the immunological synapses is a means of determining CAR-T cell behavior in an activatory context, in contrast to the anti-CD45 method mentioned herein. CAR-T cells form distinct synapses compared to wild-type T cells, which impacts their ability to properly recapitulate normal T cell function. This impairs several aspects of CAR-T activity, including signal termination, killing efficiency, and antigen sensitivity. The quality of the CAR-T synapse (i.e., similarity to the T cell synapse) correlates with clinical efficacy.

Imaging of the CAR-T cell synapse within as described herein may best be achieved using supported lipid bilayers (SLBs) as a model of the opposing antigen-presenting cell (APC) that permits high signal to noise imaging in TIRF or widefield (WF) imaging. It is also possible to image full cell-cell conjugates using 3 -dimensional imaging, however resolution will be sacrificed since the synapse will align with the z imaging plane rather than xy plane. The composition of the SLB can be modified to suit the needs of the assay, and comparison of different conditions yields further information about CAR-T behavior. The most basic SLB system for CAR-T activation would include one element to trigger CAR signaling (typically the CAR ligand or an anti-CAR affinity probe) and an adhesion ligand to allow cell capture (most commonly ICAM1 to engage LFA1 on the cell, but CD58 can also be used to engage CD2). In some embodiments, these proteins are produced recombinantly and modified to have a capture moiety that can bind to the SLB surface (typically either an oligo-histidine tail to engage Ni- bound lipid in the SLB, or a biotin group to engage avidin that is also engaged to biotinylated lipid in the SLB). Any molecule that can be modified in a manner compatible with SLB loading can be included. Typically, this will be protein ligands to cell surface receptors on the T cell - e.g. PDL1, HVEM, CD80, CD86, CD70, LIGHT, CD40, 4-1BBL, OX40L, TL1A, GITRL, CD30L, SLAM, CD48, CD58, ICAM1, CD155, CD112, CD113, Galectin9. This can also be expanded to include ligands that would normally be soluble but can be attached to the SLB surface to engage cognate receptors in a localized manner - e.g., chemokine receptor ligands, cytokine receptor ligands, Wnt receptor ligands, GPCR ligands, receptor tyrosine kinase ligands, complement receptor ligands, pattern recognition receptor ligands etc.

In some embodiments, artificial/synthetic effectors can also be included in the bilayer - e.g., small molecule inhibitors/activators/modulators, immune checkpoint inhibitors. The combination of any of these molecules can be tuned as needed, as can their absolute density on the bilayer. In a simple embodiment, the density of CAR antigen can be titrated, and the synapse features measured at different levels may be used as an indicator of the antigen sensitivity of the CAR. This could be performed in the presence and absence of other bilayer components to determine how antigen sensitivity is influenced by (e.g., by costimulation through CD80 engagement of CD28; coinhibition through PDL1 engagement of PD1; activation of distal signaling pathways such as noncanonical Wnt signaling using Foxy5 etc.).

It is also possible to include molecules in the SLB that do not engage the T cell but instead serve to capture molecules/particles released by the cell. This serves to retain those molecules at or near to the point of release within the synapse, facilitating their later detection during imaging. For example, antibodies against EV tetraspanins (CD9, CD63, CD81 etc.) or cytotoxic effectors (e.g., perforin, granzyme etc.), EV lipid-binding proteins (e.g., Tim4), membrane curvature-binding peptides etc. These capture methods can also be used in combination with any of the immobile or mobile phase activation substrates, or the non-activating substrates (e.g., a PEG surface loaded with a combination of anti-CD45 and anti-CD9 antibodies).

The morphological features listed in the previous section can all be reported within the context of the immunological synapse. These indicate the quality (i.e., similarity to ‘normal’ T cell phenotype) of the synapse - e.g., robust T cell activation is characterized by substantial spreading of the cell to form a roughly circular synapse approximately 10-30 urn in diameter. There are also several synapse-specific features that can be reported in this method. These primarily relate to the spatial organization of molecules within the synapse. A typical synapse will spontaneously organize into discrete domains (supramolecular activation clusters; SMACs) with different molecules segregated into these domains. The relative size and location of these domains, and the distribution of molecules within them is also highly informative. Poor segregation of molecules into the different SMACs is indicative of incomplete or dysfunctional activation. In the case of CAR-T cells, the location of the CAR within the synapse informs aspects of its molecular behavior and likely efficacy in vivo. For example, in normal T cells, TCRs form microclusters at the periphery of the synapse, which then migrate to the center of the contact (cSMAC) where signaling is terminated and they are internalized or released in synaptic exosomes. Most CARs do not recapitulate this behavior and instead form punctate clusters that then do not migrate - leading to inefficient signal termination and hence potential over activation of the cell. Similarly, the secretion of cytotoxic effectors by killer T cells is localized to specific regions of the synapse, which are again not frequently recreated during the activation of CAR-T cells via their CAR. Such features can be assessed with both diffraction-limited and super-resolution imaging as described herein, and the derived feature metrics may be used for cell classification.

The morphology of the contact can also be used to report different forms of cellular activation. As discussed, full synapses are characterized by circular, radially symmetrical contacts. A distinct form of contact is termed kinapse, which indicates a cell that is actively migrating while simultaneously forming an activatory contact. Kinapses are characterised by highly polarized contacts in which one end is dominated by the migratory lamellipodium, and the other by the site of TCR accumulation. Kinapses typically represent more transient contacts than kinapses (lasting minutes rather than hours in contact with a single cell), however kinapses are still a legitimate site for the release of cytotoxic effectors, and in some cases appear to be preferable as they allow a faster ‘kiss and run’ mechanism. The morphological features reported by this workflow allow the classification of contacts into synapse-like and kinapse-like. This can inform design of CAR-T cells to promote one contact type over another, and also report on the likely cell health/activity depending on the nature of the contact.

Alongside ‘global’ features of the synapse, molecular features of individual molecules or clusters of molecules can be reported to gain insights into CAR function and likely clinical efficacy. This involves the measurement of features that may vary from CAR to CAR, or across the same CAR in different cellular contexts. These may be equivalent to the features described in the herein, such as, counting, clustering, colocalization etc., however now within the context of an active synapse - e.g., how well is the CAR able to engage non-engaged ‘bystander’ TCRs to amplify signal potency in response to low levels of antigen. When CAR is able to engage bystander TCRs, in the manner that TCRs do, how many copies of recruited effectors (e.g., ZAP70) are present at each CAR/CAR cluster and how does that compare to TCR and between CARs. In some embodiments, these features serve to report how similar the CAR is behaving to TCR at the molecular level, and how signaling derived from the CAR is propagating into the cell, both of which may be informative for CAR design.

Another feature of the immunological synapse is the force generated by the cell on the activatory substrate. During T cell activation, active cytoskeletal rearrangements and Brownian motion leads the generation of pN-scale forces within the synapse, with significant impacts on cellular response. The strength, nature, and location of these forces is related to the potency and quality of cellular activation, and so metrics relating to such forces can provide insights into CAR-T cell responses and potential efficacy. Within this workflow, force generation can be achieved in two exemplary ways. The first involves the introduction of force-sensitive spectroscopic rulers into the bilayer, conjugated to molecules that engage the T cell surface. These molecules change their fluorescence intensity and/or spectral properties depending on the force applied, and so can be used for molecular tension fluorescence microscopy. Alternatively, cells can be activated on a semi-solid gel substrate containing embedded fluorescent beads. Traction force microscopy may be used, where mechanical forces applied by the cell on the substrate deform the gel and so move the beads in the direction of the pulling force. Both molecular tension fluorescence microscopy and traction force microscopy can be used in conjunction with SMLM-based methods or diffraction-limited imaging of the other targets as described herein.

Although the CAR-T synapse would typically be induced through engagement of the CAR, it is also possible to induce cell activation through engagement of the endogenous TCR that is typically still present in the cell after CAR transduction. TCR can be activated using anti-TCR or anti-CD3 affinity probes, cognate pMHC, or superantigen (e.g., SEE, SEB etc.) immobilized on the activation surface. This can be performed in systems that also engage CAR, or that engage TCR alone. Activating CAR-T cells through the TCR provides insights into the inherent activation potential of the cells rather than the specific details of CAR signaling - e.g., exhausted cells may activate poorly through TCR or CAR, as will stressed, senescent or non-viable cells. The inventors have recognized and appreciated that synaptic features may be used to report cell health/status in advance of transduction with CAR in order to assess the quality of the T cell pool, and by extension the likelihood of success of the CAR approach. Activation through both the CAR and TCR simultaneously may be used to obtain insights into how combinatorial signaling from both receptors influence cell behavior. This could provide valuable insights into the development of any therapies that involve targeting of both CAR and TCR against cancer cells - e.g., combining CAR-T therapy with immune mobilizing monoclonal TCRs against cancer (immTAC) treatment (Immunocore® - involves cross-linking TCRs to cancer targets using soluble, bifunctional linker molecules).

VI. Relevance to CAR-T function

The feature metrics listed above form a very diverse set of parameters that can be derived in isolation or in combination, and then used as part of the cell classification workflow described in the broader disclosure. The overall utility of this is to allow the connections and relationships between nanoscale or microscale organization to cell activity and clinical efficacy to be determined. In practice, metrics reporting marker density, nanoscale organization, and microscale organization may be combined with metrics relating to overall cell morphology to create cellular classifications that correlates with cellular behavior, which in turn allows correlation of different behavior modes to functional responses (e.g., cytotoxicity) and subsequent clinical performance. The main characteristics that can be assessed in preclinical models and then correlated with these classifications are treatment efficacy (reported as patient survival and remission rates), safety, and side effects. The fact that the cellular classification will be influenced by subtle changes in CAR design (e.g., leading to differences in expression, nanoscale distribution, signaling conformation, interaction with other proteins etc.) means that this workflow offers an integrated solution for linking early CAR development with likely clinical outcomes, and hence inform early design work.

VIE Example use cases

The incorporation of single-molecule characterization data into multidimensional cell analyses has the substantial advantage of reporting several dimensions that are not accessible to diffraction-limited imaging methods. Combined with the fact that biomarker detection in this process is inherently single-molecule-sensitive, cellular characterization in the manner described herein can be applied to a wide range of biological and biomedical areas. These include, but are not limited to, the following examples.

Adoptive cell transfer (ACT) is a rapidly growing field of therapeutics in a number of disease areas, most notably cancer. ACT represents a range of different cellular therapies in which live effector cells are transfused into a patient to perform a desired function. Most commonly, these cells are directly derived from the patient and modified in vitro to direct their function upon readministration. The most widely used examples of this are chimeric antigen receptors (CARs), which are artificially designed antigen receptors consisting of an extracellular domain capable of specifically recognizing a given target molecule, and an intracellular domain consisting of signaling motifs capable of directing a given cellular outcome in response to said target. This typically consists of a single chain variable fragment (scFV) as the extracellular domain, and signaling motifs derived from the immunoreceptor tyrosine-based activation motifs (IT AMs) of CD3(" and costimulatory receptors such as CD28 or 4-1BB. The most commonly used cells into which CARs are introduced are CD4+ and CD8+ aPT cells, however more innate-like T cells (such as NKT or ydT cells) and other lymphocyte (e.g., NK cells) or myeloid (e.g., macrophages) are also undergoing active development as clinical therapies. CAR-T therapy alone is beginning to play a substantial role in treatment for many cancers, with five therapies currently approved by the FDA for treatment of hematological malignancies such as acute lymphoblastic leukemia, diffuse large B cell lymphoma, and multiple myeloma. Alongside CAR-T and other CAR therapies, active development is ongoing for many other forms of ACT, including tumorinfiltrating lymphocyte (TIL) therapy, and engineered T cell receptor (eTCR) therapy.

Profiling the status of cells used for ACT is of key importance during the development of new approaches and the administration of therapies to patients. This is most widely performed by measuring the abundance of cell-surface molecules as this the most straightforward in terms of experimental design yet is still highly informative due to the highly responsive nature of lymphocyte and myeloid cell surface proteomes. Such experiments can report a wide range of features, including the expression of artificial constructs, markers of health, markers of activation, markers of population/identity, and many others. Profiling of intracellular molecules/features (e.g., specific nucleic acid sequences, metabolites, molecules undergoing secretion, or organelle morphologies associated with cell health etc.) is also commonly undertaken. The vast majority of experiments done with single-cell resolution are performed using flow cytometry. There are currently no methods available for the routine reporting of the spatial organization of biomarkers of interest in ACT. Within ACT, this analysis has many potential applications, including the following:

T cell exhaustion. A major determinant of the efficacy of T cell-based ACT therapy is the extent to which transferred cells exhibit an exhausted phenotype, which is characterized by poor effector function and a greatly reduced barrier to inhibition by immune checkpoint molecules. Exhaustion is a complex phenotype that can be induced by several factors, most notably by chronic exposure to stimulatory antigen as occurs in chronic conditions such as cancer. Substantial development is ongoing to develop CAR systems that can minimize exhaustion or counteract its effects. For both development purposes, and the profding of exhaustion phenotype at the point of administration to patients, reporting the exhaustion status of CAR-T cells has substantial value.

Current methods of exhaustion profiling using flow cytometry or bulk analyses are subject to the limitations of interpretation discussed herein. Exhaustion is a complex phenotype that cannot be detected simply by the presence or absence of one marker, instead requiring the detection of multiple markers in combination. The most informative such markers may be inhibitory receptors such as PD1, CTLA-4, TIM3, LAG3, and TIGIT, in part because these are also effectors of exhaustion since signaling from these receptors strongly inhibits cell activation in response to antigen. Increased expression of just one such marker does not in itself denote exhaustion, however the concurrent increase of multiple markers within a population indicates a shift towards an exhausted phenotype. The ability to detect these markers at levels of expression below the sensitivity limit of flow cytometry allows such a shift to be reported at an earlier stage, potentially allowing intervention to rescue the phenotype from exhaustion. Similarly, more sensitive detection allows for greater precision in biomarker quantification, such that moderate changes in expression can be reported with far greater confidence. This means that not only can the presence of these markers be reported earlier in the exhaustion process, but also that any increases in their densities over time can be more readily detected. In combination, this means that the status of a T cell population can be reported with a much greater sensitivity for cells undergoing early signs of exhaustion. Since the major makers of exhaustion are also effectors of T cell inhibition, profiling their spatial organization is also highly informative. The extent of clustering of these molecules may be correlated with their extent of phosphorylation (since phosphorylation of tyrosine-based signaling motifs typically leads to receptor clustering), and so report the strength of underlying inhibitor signaling. The relative organization of such molecules to antigen receptors such as CARs or eTCRs is also informative. Many such receptors mediate effector function by recruiting inhibitory phosphatases to their cytosolic signaling domains (e.g., SHP1 recruitment by PD1), which are then only able to exert inhibitory effects on receptors located within the dynamic diffusion radius of the signaling domain. As a result, inhibitory receptors that cluster and colocalize with activatory receptors have a stronger inhibitory effect than those that do not. By reporting spatial organization, it is therefore possible to obtain biologically meaningful insights beyond simple overall density.

This can impact a range of ACT approaches at several points of development and administration. For example: 1. CAR-T cells can be sampled and profded at multiple points during the expansion process in order to predict the point at which exhaustion will occur and so stop expansion before this point; 2. T cells from a patient can be profded for exhausted and precursor exhausted cells prior to transduction with CAR or eTCR constructs in order to inform the likelihood of successful expansion and so triage resources accordingly; 3. new CARs undergoing development can be characterized according to their capacity to drive exhaustion phenotypes in transduced cells or to colocalize with coinhibitory receptors, and so inform development of CARs that minimize exhaustion.

Antigen-receptor expression. The majority of ACT approaches involve the expression of engineered antigen receptors such as CARs or eTCRs. The efficacy of many such receptors is still high even at low levels of expression due to the highly amplified nature of signaling upon their engagement by cognate ligand. For this reason, cells expressing levels of antigen receptors below the detection limit of flow cytometry and other methods may still contribute to tumor-clearance in patients. It is therefore informative to report the bona fide expression of such receptors on these cells so that the full functional population can be identified. The spatial organization of these molecules is also important to assess as it directly impacts the potential for activation. Highly clustered IT AM-bearing receptors typically have a lower threshold for triggering than nonclustered receptors due to the recruitment of secondary kinases (e.g., ZAP70 in the case of most CARs and TCR/eTCR), which then transphosphorylate other receptors in the same cluster. The inherent capacity of different engineered receptors to drive clustering or induce clustering in response to sub-threshold ‘tonic’ signaling will vary according to receptor architecture, and so reporting this phenotype is of inherent utility in understanding and refining receptor design. This can be combined with staining for recruited effector molecules (e.g., ZAP70) and/or post- translational modifications on the receptor (e.g., pTyr for CAR) or recruited effectors (e.g., phospho-ZAP70) to further report the activation status and potential of the receptors of interest. In such cases, colocalization of these secondary receptors with the receptor of interest would be a major metric to be assessed.

This has utility in both cell manufacture (e.g., what is the true fraction of cells with the correct density and spatial profile for receptor(s) of interest) and development (e.g., what is the typical express! on/spatial profile of a particular antigen receptor). This is applicable to a wide range of cell types used in ACT, including T cells, NK cells, and macrophages. NK cell responsiveness. NK cells are currently being explored as cytotoxic cells to be used in ACT, with some potential benefits over T cells due to their lack of clonality. The surface profile of NK cells has a significant impact on their activation and response potential due to complex interplay between different receptors at the cell surface. For example, activation of NK cells through the receptor NKG2D leads to reduction in responsiveness to antibody-opsonized targets due to the protease-dependent shedding of the CD16 Fc receptor extracellular domain. The interplay between such receptors therefore strongly impacts the potential for NK cell activation. For example, if expression of a CAR causes a reduction in CD 16 density then the potential for simultaneous CAR- and ADCC-mediated therapy is poorer. Correlation of such an effect with endogenous spatial behavior of the CAR in question can then be used to refine the design of novel NK-expressed CARs to minimize such effects. Accurate profiling of the density of such receptors at the NK cell surface to high sensitivity therefore has key potential to report the responsiveness of such cells upon administration.

Macrophage receptor clustering. Macrophages are some of the other leukocytes being developed for ACT, as so-called CAR-M therapy. Macrophages are typically activated to phagocytose targets by activation of Fc receptors (FcR), complement receptors, or other innate immune receptors. The polarized signaling of such proteins upon engagement of a target induces cytoskeletal reorganization to allow engulfment and killing. As with many other such receptors, the activation process involves signal amplification via clustering. Indeed, FcRs can be activated through clustering alone. CARs introduced into macrophages may be able to increase their potency through the co-clustering with such endogenous receptors to co-opt their signaling upon target engagement, particularly since such proteins will typically be far more abundant than ectopically expressed antigen receptors. Characterization of the relative clustering of CARs and other stimulatory receptors will therefore provide insights into the likely responsiveness of the cells in vivo. In particular, this may inform the likely efficacy of combinatorial CAR-M monoclonal-Ab therapy, as the synergy between FcR-mediated responses to target-bound mAbs and CAR-mediated responses to tumor antigens will be heavily influenced by the relative nanoscale organization.

Tumor profiling. Alongside the profiling of engineered effector cells, the characterization of tumor cells is also a highly informative embodiment of this workflow. Ultrasensitive detection of CAR-targeted antigens by dSTORM is clinically relevant because CAR-T cell responsiveness is still robust at densities of target antigen well below the sensitivity of other methods. Molecular counting in this way can be combined with the characterization of biomarker spatial organization to provide yet further insights. The spatial nature of the targets of antigen receptors and other cellsurface molecules will have a profound impact on clinical response. For example, highly clustered antigen may induce more potent responses from its cognate receptor than less densely clustered antigen. Hence the combination of overall antigen density and antigen clustering may inform which therapy a patient should receive - e.g., a low-density, non-clustered antigen may warrant an inherently more highly clustered CAR to compensate for the low potency. This can be extended to look at coreceptor ligands on the tumor cell surface - e.g., if the target antigen clusters tightly with PDL1 then CAR-T cells that have high PD1 will have a poorer response (due to the consequent colocalization of CAR and PD1 that will occur upon engagement) than if the target and PDL1 are segregated. By building up a detailed profde of the molecular organization of both target and effector cells, treatment regimens can be better refined and optimized.

Tumor-infiltrating lymphocyte (TIL) profiling. The abundance and identity of TILs has a profound impact on clinical outcome in many solid cancers. Tumor restriction or clearance is typically most successful in cases with a high abundance of active effector cytotoxic cells, and least successful in cases with a high abundance or exhausted/anergic cells and/or regulatory T cells. The TIL profile for a given tumor is typically assessed at the level of whole cell identity - i.e., fraction of T cells, B cells, NK cells etc. and subtypes thereof (memory, effector, regulatory etc.) - and there is no characterization of the nanoscale phenotype as discussed in this disclosure. Profiling TIL populations in the manner described here would allow more subtle clinical correlates to be identified - e.g., correlations between the degree of TCR-PD1 colocalization with survival etc. Such information with provide key insights into TIL biology that can inform the development of TIL-focused therapies (e.g., immune checkpoint blockade) and be of use when assessing patient prognosis (i.e., do their native TIL populations demonstrate molecular characteristics the correlate with good or poor clinical outcomes).

Statements

The following numbered paragraphs set out statements forming part of the present disclosure:

AL A method for measuring single molecule emission from a cell, the method comprising: applying a fixation agent to the cell while the cell is in a native configuration to produce a fixed cell; capturing the fixed cell on a surface of a sample substrate; and acquiring an imaging sequence of single molecule emission from the fixed cell. A2. The method of Al, wherein capturing the fixed cell on the surface of the sample substrate comprises flattening a cell membrane of the fixed cell on the surface.

A3. The method of A2, further comprising: determining, using at least one first image from the imaging sequence, an area corresponding to the fixed cell membrane; and determining, using at least one second image from the imaging sequence, a plurality of proteins in the cell membrane and/or in the fixed cell.

A4. The method of A3, wherein the at least one first image and the at least one second image are fluorescence images.

A5. The method of A3, wherein the at least one first image is a white light image and the at least one second image is a fluorescence image.

A6. The method of A3, further comprising determining, using the at least one second image, a location of an individual emitter associated with a protein on or in the fixed cell, wherein the location is determined to within a 100 nm resolution.

A7. The method of A6, further comprising determining, using the location of the individual emitter, the spatial organization of a molecule associated with the individual emitter relative to the plurality of proteins in the cell membrane and/or in the fixed cell.

A8. The method of A7, wherein determining the spatial organization of a molecule associated with the individual emitter, comprises determining the spatial organization of the number of the cell structures using a hierarchical clustering of a plurality of fluorescence signals.

A9. The method of Al, wherein capturing the fixed cell comprises capturing the fixed cell using an agnostic capture surface.

A10. The method of Al, wherein capturing the fixed cell comprises capturing the fixed cell using a capture surface configured to bind to specific cells and/or proteins.

All. The method of Al , wherein the fixed cell is captured on a front surface of the sample substrate, and wherein acquiring the imaging sequence comprises acquiring the image sequence using a microscope configured to illuminate a back side of the sample substrate at an angle such that the illumination light undergoes total internal reflection.

Bl. An apparatus configured to provide a method for measuring single molecule emission from a cell, the apparatus comprising at least one processor in communication with memory and a set of additional processing resources, the processor being configured to execute instructions stored in the memory that cause the apparatus to: apply a fixation agent to the cell while the cell is in a native configuration to produce a fixed cell; capture the fixed cell on a surface of a sample substrate; and acquire an imaging sequence of single molecule emission from the fixed cell.

B2. The apparatus of Bl, wherein capturing the fixed cell on the surface of the sample substrate comprises flattening a cell membrane of the fixed cell on the surface.

B3. The apparatus of B2, wherein the instructions are further configured to cause the apparatus to: determine, using at least one first image from the imaging sequence, an area corresponding to the fixed cell membrane; and determine, using at least one second image from the imaging sequence, a plurality of proteins in the cell membrane and/or in the fixed cell.

B4. The apparatus of B2, wherein the instructions are further configured to cause the apparatus to: determine, using the at least one second image, a location of an individual emitter associated with a protein on or in the cell, wherein the location is determined to within a 100 nm resolution; and determine, using the location of the individual emitter, the spatial organization of a molecule associated with the individual emitter relative to the plurality of proteins in the cell membrane and/or in the fixed cell.

Cl . At least one non-transitory computer-readable storage medium encoded with a plurality of computer-executable instructions that, when executed, perform a method for measuring single molecule emission from a cell, the method comprising: applying a fixation agent to the cell while the cell is in a native configuration to produce a fixed cell; capturing the fixed cell on a surface of a sample substrate; and acquiring an imaging sequence of single molecule emission from the fixed cell.

C2. The at least one non-transitory computer-readable storage medium of Cl, wherein capturing the fixed cell on the surface of the sample substrate comprises flattening a cell membrane of the fixed cell on the surface.

C3. The at least one non-transitory computer-readable storage medium of C2, wherein the method further comprises: determining, using at least one first image from the imaging sequence, an area corresponding to the fixed cell membrane; and determining, using at least one second image from the imaging sequence, a plurality of proteins in the cell membrane and/or in the fixed cell. C4. The at least one non-transitory computer-readable storage medium of C3, wherein the method further comprises: determining, using the at least one second image, a location of an individual emitter associated with a protein on or in the cell, wherein the location is determined to within a 100 nm resolution; and determining, using the location of the individual emitter, the spatial organization of a molecule associated with the individual emitter relative to the plurality of proteins in the cell membrane and/or in the fixed cell.

C5. The at least one non-transitory computer-readable storage medium of C4, wherein determining the spatial organization of a molecule associated with the individual emitter, comprises determining the spatial organization of the number of the cell structures using a hierarchical clustering of a plurality of fluorescence signals.

DI. A system for measuring a single molecule emission from a cell, the system comprising: at least one computer hardware processor; and at least one non-transitory computer-readable storage medium storing processorexecutable instructions, that, when executed by the at least one computer hardware processor, cause the at least one computer hardware processor to perform a method for measuring single molecule emission from a cell, the method comprising: applying a fixation agent to the cell while the cell is in a native configuration to produce a fixed cell; capturing the fixed cell on a surface of sample substrate; and acquiring an imaging sequence of single molecule emission from the fixed cell.

D2. The system of DI, further comprising: a plurality of microfluidic components configured to the fixation agent to the cell; and a plurality of optical components configured to acquire single molecule emission from the fixed cell.

D3. The system of DI, wherein the plurality of optical components are configured to illuminate a back side of the sample substrate at an angle such that the illumination light undergoes total internal reflection.

D4. The system of DI, wherein the method for measuring single molecule emission further comprises: determining, using at least one first image from the imaging sequence, an area corresponding to the fixed cell membrane; and determining, using at least one second image from the imaging sequence, a plurality of proteins in the cell membrane and/or in the fixed cell.

D5. The system of DI, wherein the method for measuring single molecule emission further comprises further comprises determining, using at least one image, a location of an individual emitter associated with a protein on or in the fixed cell, wherein the location is determined to within a 100 nm resolution.

Additional Comments

When techniques described herein are embodied as computer-executable instructions, these computer-executable instructions may be implemented in any suitable manner, including as a number of functional facilities, each providing one or more operations to complete execution of algorithms operating according to these techniques. A “functional facility,” however instantiated, is a structural component of a computer system that, when integrated with and executed by one or more computers, causes the one or more computers to perform a specific operational role. A functional facility may be a portion of or an entire software element. For example, a functional facility may be implemented as a function of a process, or as a discrete process, or as any other suitable unit of processing. If techniques described herein are implemented as multiple functional facilities, each functional facility may be implemented in its own way; all need not be implemented the same way. Additionally, these functional facilities may be executed in parallel and/or serially, as appropriate, and may pass information between one another using a shared memory on the computer(s) on which they are executing, using a message passing protocol, or in any other suitable way.

Generally, functional facilities include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Typically, the functionality of the functional facilities may be combined or distributed as desired in the systems in which they operate. In some implementations, one or more functional facilities carrying out techniques herein may together form a complete software package. These functional facilities may, in alternative embodiments, be adapted to interact with other, unrelated functional facilities and/or processes, to implement a software program application.

Cell classification techniques, as described herein, may be implemented as a functional facilities have been described herein for carrying out one or more of the tasks or processes described herein. It should be appreciated, though, that embodiments are not limited to being implemented in any specific number, division, or type of functional facilities. In some implementations, all functionality may be implemented in a single functional facility. It should also be appreciated that, in some implementations, some of the functional facilities described herein may be implemented together with or separately from others (i.e., as a single unit or separate units), or some of these functional facilities may not be implemented.

Computer-executable instructions implementing the techniques described herein (when implemented as one or more functional facilities or in any other manner) may, in some embodiments, be encoded on one or more computer-readable media to provide functionality to the media. Computer-readable media include magnetic media such as a hard disk drive, optical media such as a Compact Disk (CD) or a Digital Versatile Disk (DVD), a persistent or non- persistent solid-state memory (e.g., Flash memory, Magnetic RAM, etc.), or any other suitable storage media. Such a computer-readable medium may be implemented in any suitable manner. As used herein, “computer-readable media” (also called “computer-readable storage media”) refers to tangible storage media. Tangible storage media are non-transitory and have at least one physical, structural component. In a “computer-readable medium,” as used herein, at least one physical, structural component has at least one physical property that may be altered in some way during a process of creating the medium with embedded information, a process of recording information thereon, or any other process of encoding the medium with information. For example, a magnetization state of a portion of a physical structure of a computer-readable medium may be altered during a recording process.

Further, some techniques described above comprise acts of storing information (e.g., data and/or instructions) in certain ways for use by these techniques. In some implementations of these techniques — such as implementations where the techniques are implemented as computerexecutable instructions — the information may be encoded on a computer-readable storage media. Where specific structures are described herein as advantageous formats in which to store this information, these structures may be used to impart a physical organization of the information when encoded on the storage medium. These advantageous structures may then provide functionality to the storage medium by affecting operations of one or more processors interacting with the information; for example, by increasing the efficiency of computer operations performed by the processor(s). In some, but not all, implementations in which the techniques may be embodied as computerexecutable instructions, these instructions may be executed on one or more suitable computing device(s) operating in any suitable computer system, or one or more computing devices (or one or more processors of one or more computing devices) may be programmed to execute the computer-executable instructions. A computing device or processor may be programmed to execute instructions when the instructions are stored in a manner accessible to the computing device or processor, such as in a data store (e.g., an on-chip cache or instruction register, a computer-readable storage medium accessible via a bus, a computer-readable storage medium accessible via one or more networks and accessible by the device/processor, etc.). Functional facilities comprising these computer-executable instructions may be integrated with and direct the operation of a single multi-purpose programmable digital computing device, a coordinated system of two or more multi-purpose computing device sharing processing power and jointly carrying out the techniques described herein, a single computing device or coordinated system of computing device (co-located or geographically distributed) dedicated to executing the techniques described herein, one or more Field-Programmable Gate Arrays (FPGAs) for carrying out the techniques described herein, or any other suitable system.

FIG. 4 illustrates one exemplary implementation of a computing device in the form of computing device 400 that may be used in a system implementing techniques described herein, although others are possible. It should be appreciated that FIG. 4 is intended neither to be a description of necessary components for a computing device to operate as cell classification facility, in accordance with the techniques described herein, nor a comprehensive depiction.

Computing device 400 may include at least one processor 402, a network adapter 404, and a nonvolatile computer-readable storage media 406. Computing device 400 may be, for example a desktop or laptop personal computer, a personal digital assistant, a smart mobile phone, HOT equipment, or any other suitable computing device. Network adapter 404 may be any suitable hardware and/or software to enable the computing device 400 to communicate through wired and/or wireless connections with any other suitable computing device over any suitable computing network and using any suitable networking protocol, as described herein. The computing network may include switches, routers, gateways, access points, and/or other networking equipment as well as any suitable wired and/or wireless communication medium or media for exchanging data between two or more computers, including the Internet. Non-volatile computer readable storage media 406 may be adapted to store data to be processed and/or instructions to be executed by processor 402. Processor 402 enables processing of data and execution of instructions. The data instructions may be stored on the computer-readable storage media 406. The processor 402 may control writing data to and reading data from the non-volatile computer-readable storage media 406 and memory 410 in any suitable manner, as the aspects of the disclosure provided herein are not limited in this respect.

The data and instructions stored on computer-readable storage media 406 may include computerexecutable instructions implementing techniques which operate according to the techniques described herein. In the example of FIG. 4, non-volatile computer-readable storage media 406 stores computer-executable instructions implementing various facilities and storing various information as described above. Non-volatile computer-readable storage media 406 functional facility for performing cell classification techniques, in accordance with some embodiments described herein.

While not illustrated in FIG. 4, a computing device may additionally have one or more components and peripherals, including input and output devices. These devices can be used, among other things to present a user interface. Examples, of output devices that can be used to provide a user interface include printers or display screens for visual presentation of output and speakers or other sound generating devices for audible presentation of output. Example of input devices that can be used for a user interface include keyboards, and pointing devices, such as mice, touch pads, and digitizing tablets. As another example, a computing device may receive input information through speech recognition or in other audible format.

Techniques operating according to the principles described herein may be implemented in any suitable manner. The processing and decision blocks of the flowcharts above represent steps and acts that may be included in algorithms that carry out these various processes. Algorithms derived from these processes may be implemented as software integrated with and directing the operation of one or more single- or multi-purpose processors, may be implemented as functionally- equivalent circuits such as a Digital Signal Processing (DSP) circuit or an Application-Specific Integrated Circuit (ASIC), or may be implemented in any other suitable manner. It should be appreciated that the flowcharts included herein do not depict the syntax or operation of any particular circuit or of any particular programming language or type of programming language. Rather, the flowcharts illustrate the functional information one skilled in the art may use to fabricate circuits or to implement computer software algorithms to perform the processing of a particular apparatus carrying out the types of techniques described herein. It should also be appreciated that, unless otherwise indicated herein, the particular sequence of steps and/or acts described in each flowchart is merely illustrative of the algorithms that may be implemented and can be varied in implementations and embodiments of the principles described herein.

Accordingly, in some embodiments, the techniques described herein may be embodied in computer-executable instructions implemented as software, including as application software, system software, firmware, middleware, embedded code, or any other suitable type of computer code. Such computer-executable instructions may be written using any of a number of suitable programming languages and/or programming or scripting tools, and also may be compiled as executable machine language code or intermediate code that is executed on a framework or virtual machine.

Embodiments have been described where the techniques are implemented in circuitry and/or computer-executable instructions. It should be appreciated that some embodiments may be in the form of a method, of which at least one example has been provided. The acts performed as part of the method may be ordered in any suitable way. Accordingly, embodiments may be constructed in which acts are performed in an order different than illustrated, which may include performing some acts simultaneously, even though shown as sequential acts in illustrative embodiments.

Various aspects of the embodiments described above may be used alone, in combination, or in a variety of arrangements not specifically discussed in the embodiments described in the foregoing and is therefore not limited in its application to the details and arrangement of components set forth in the foregoing description or illustrated in the drawings. For example, aspects described in one embodiment may be combined in any manner with aspects described in other embodiments.

The phrase “and/or,” as used herein in the specification and in the claims, should be understood to mean “either or both,” of the elements so conjoined, i.e., elements that are conjunctively present in some cases and disjunctively present in other cases. Multiple elements listed with “and/or” should be construed in the same fashion, i.e., “one or more” of the elements so conjoined. Other elements may optionally be present other than the elements specifically identified by the “and/or” clause, whether related or unrelated to those elements specifically identified. Thus, as a non-limiting example, a reference to “A and/or B,” when used in conjunction with open-ended language such as “comprising” can refer, in one embodiment, to A only (optionally including elements other than B); in another embodiment, to B only (optionally including elements other than A); in yet another embodiment, to both A and B (optionally including other elements); etc.

The indefinite articles “a” and “an,” as used herein in the specification and in the claims, unless clearly indicated to the contrary, should be understood to mean “at least one.”

As used herein in the specification and in the claims, the phrase, “at least one,” in reference to a list of one or more elements, should be understood to mean at least one element selected from any one or more of the elements in the list of elements, but not necessarily including at least one of each and every element specifically listed within the list of elements and not excluding any combinations of elements in the list of elements. This definition also allows that elements may optionally be present other than the elements specifically identified within the list of elements to which the phrase “at least one” refers, whether related or unrelated to those elements specifically identified. Thus, as a non-limiting example, “at least one of A and B” (or, equivalently, “at least one of A or B,” or, equivalently, “at least one of A and/or B”) can refer, in one embodiment, to at least one, optionally including more than one, A, with no B present (and optionally including elements other than B); in another embodiment, to at least one, optionally including more than one, B, with no A present (and optionally including elements other than A); in yet another embodiment, to at least one, optionally including more than one, A, ,and at least one, optionally including more than one, B (and optionally including other elements); etc.

Use of ordinal terms such as “first,” “second,” “third,” etc., in the claims to modify a claim element does not by itself connote any priority, precedence, or order of one claim element over another or the temporal order in which acts of a method are performed, but are used merely as labels to distinguish one claim element having a certain name from another element having a same name (but for use of the ordinal term) to distinguish the claim elements.

Also, the phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting. The use of “including,” “comprising,” “having,” “containing,” “involving,” and variations thereof herein, is meant to encompass the items listed thereafter and equivalents thereof as well as additional items. All definitions, as defined and used herein, should be understood to control over dictionary definitions, definitions in documents incorporated by reference, and/or ordinary meanings of the defined terms.

The word “exemplary” is used herein to mean serving as an example, instance, or illustration.

Any embodiment, implementation, process, feature, etc. described herein as exemplary should therefore be understood to be an illustrative example and should not be understood to be a preferred or advantageous example unless otherwise indicated.

Having thus described several aspects of at least one embodiment, it is to be appreciated that various alterations, modifications, and improvements will readily occur to those skilled in the art. Such alterations, modifications, and improvements are intended to be part of this disclosure and are intended to be within the spirit and scope of the principles described herein. Accordingly, the foregoing description and drawings are by way of example only.

Claims

CLAIMS:

1. A substrate comprising an imaging surface suitable for immobilizing and imaging leukocytes, the imaging surface having attached thereto:

2. A substrate according to claim 1, wherein the CD45 capture molecule is an anti-CD45 antibody.

3. A substrate according to claim 2, wherein the anti-CD45 antibody targets a binding epitope on CD45 located distally to the cell membrane, preferably wherein the anti-CD45 antibody targets a binding epitope within or including the first immunoglobulin (Ig) domain of CD45 (as determined from the N-terminus) or the common domain of the mucin-like domain of CD45.

4. A substrate according to any one of the preceding claims, wherein the CD45 capture molecule bears an anchor moiety, and the secondary capture molecule likewise bears an anchor moiety, wherein the interaction between the CD45 capture molecule and secondary capture molecule occurs via a mediating compound having multiple binding moieties suitable for binding to said anchor moieties.

5. A substrate according to claim 2 or 3, wherein the CD45 capture molecule is a biotinylated anti-CD45 antibody, and the secondary capture molecule is PEG linked to biotin, which is then coated with neutravidin, which in turn captures the biotinylated anti-CD45 antibody.

6. A substrate according to claim 5, wherein the passivation reagent is PEG.

7. A substrate according to claim 5 or 6, wherein the passivation reagent is non-biotinylated PEG, and the ratio of biotinylated-PEG to non-biotinylated PEG is no more than 0.5:1.

8. A substrate according to any one of claims 1 to 7, wherein the imaging surface further comprises competitor molecules which attach to the secondary capture molecule.

9. A substrate according to claim 8 when dependent on any one of claims 5 to 7, wherein the competitor molecule is free biotin.

10. A substrate according to claim 1, comprising

- a passivation reagent, being non-biotinylated PEG attached to the imaging surface;

- biotinylated PEG attached to the imaging surface;

- neutravidin attached to the biotinylated PEG; and

- biotinylated anti-CD45 antibody, attached to said neutravidin.

11. A kit of parts for making a substrate according to any one of claims 1 to 10.

12. A kit of parts according to claim 11, comprising:

- said passivation reagent, being non-biotinylated PEG;

- said secondary capture molecule, being biotinylated PEG

- biotinylated anti-CD45 antibody; and

- neutravidin.

13. A method of preparing leukocytes for imaging, comprising:

(i) providing an imaging surface, having attached thereto: a. a passivation reagent, for preventing non-specific binding to the imaging surface; and b. a CD45 capture molecule, the CD45 capture molecule being tethered to the imaging surface via a secondary capture molecule;

(iii) fixing the leukocytes by applying a fixation agent.

14. A method according to claim 13, wherein the leukocyte is a T cell or a CAR-T cell.

15. A method according to claim 13 or 14, wherein the imaging surface is as defined in any one of claims 1 to 10.

16. A method according to any one of claims 13 to 15, wherein step (ii) is carried out with cells suspended in growth medium.

17. A method according to any one of claims 13 to 16, wherein step (ii) is followed by step (ii-A), comprising flushing the imaging surface with a flushing liquid.

18. A method according to claim 17, wherein the flushing liquid is growth medium, preferably serum -free growth medium.

19. A method according to any one of claims 13 to 18, wherein the time between the start of step (ii) and the initiation of step (iii) is at least at least 5 minutes.

20. A method according to any one of claims 13 to 19, wherein the method comprises an additional step (iv) of staining the leukocytes with one or more fluorescent stains or fluorescent probes.

21. A method of carrying out fluorescence imaging of leukocytes, comprising preparing leukocytes for imaging according to claim 20, and carrying out fluorescence imaging of biomarkers on the leukocytes.

22. A method according to claim 21, wherein the fluorescence imaging is single molecule localization microscopy to obtain spatial coordinates of said biomarkers.

23. A method according to claim 22, wherein the fluorescence imaging is dSTORM, PALM or PAINT.

24. A method according to any one of claims 21 to 23, wherein the leukocytes are T cells, preferably CAR-T cells.

25. A method of identifying whether a sample of T cells from a patient is suitable for use as therapeutic cells in CAR-T cell therapy, comprising imaging the sample of T cells using a method according to claim 22 or 23, to obtain spatial coordinates of a biomarker on the T cells; detecting boundaries of the plurality of cells; constructing a sample feature vector based on the obtained spatial coordinates and the detected boundaries; providing reference data, wherein the reference data comprises one or more reference feature vectors obtained for reference cells, wherein the reference cells are CAR-T cells from patients with a known therapeutic outcome; and carrying out data analysis, comprising comparing the sample feature vector with said reference feature vector(s), and determining the similarity of the plurality of cells to the reference cells, wherein a greater degree of similarity is indicative of a greater suitability for use in CAR-T cell therapy.