Technical Field

This invention concerns a coating for capture of bioanalytes comprising hyaluronic acid (HA) that is end-point attached to the surface of a medical sampling device. Moreover, this invention concerns a medical sampling device for capture of circulating tumor cells (CTCs) for subsequent analysis. More in particular, this invention concerns a medical sampling device, provided with a coating resulting in the enhanced capture of CTCs which can then be released for analysis and diagnosis.

Background Art

Devices are known for the capture of bioanalytes wherein the object to be analyzed is bioactive and is selected from the group comprising macromolecules, polynucleotides, RNA, DNA, proteins, marker proteins, lipoproteins, polypeptides, antibodies, autoantibodies, hormones, antigens, cells, CD44+ cells, viruses, bacterial cells, parasites, fungus cells, tumor cells, stem cells and/or cells that originate from a fetus during pregnancy, or parts thereof. Of particular interest are devices for the capture of circulating tumor cells.

Chemotherapies have become much more advanced over the past two decades. Whereas older 'generic' chemotherapies simply kill all cells in the body that are growing quickly (resulting in damage to healthy tissues), modern targeted chemotherapies are designed to only affect specific (cancer) cells and minimize collateral damage. Although targeted chemotherapies have provided a major improvement in cancer care, their true potential remains unrealized. Because cancer cells are continuously mutating, they can adapt and become resistant against a targeted chemotherapy. This ultimately results in the targeted chemotherapy becoming ineffective. Considering that cancer is often fatal and targeted chemotherapies are expensive, there is a great need for tools that allow for the tracking of when and how the tumor cells have become resistant, so that the therapy can be adjusted accordingly.

Currently the most widely accepted way to track tumor mutations is to take a biopsy, a direct tissue sample of the tumor. This is an invasive procedure, which is inherently painful, carries a risk of locally spreading the disease, and even when performed well only gives a reflection of the precise location sampled at that particular time. Due to patient considerations successive biopsies of all tumors are rarely performed. Liquid biopsies sample the blood instead of the tumor, and are the most promising tools for continuously tracking tumor mutations because of their relative ease and safety. There are many different liquid biopsy approaches, including the isolation of circulating tumor cells (CTCs), tumor derived vesicles (e.g. exosomes) and circulating tumor DNA (ctDNA). Of all these approaches, CTCs provide the most comprehensive and in depth information on tumor resistance, because they contain a complete profile of DNA, RNA and proteins. However, CTCs are extremely rare in the blood: typical blood samples average less than 1 CTC per milliliter of blood, among billions of other cells. Many more cells (100+) are needed to track tumor resistance reliably, which is why there is a need for technologies that can efficiently isolate CTCs from the blood. The present invention therefore concerns the selective capture of CTCs.

The application of optical fiber, catheter or wire-based devices, which are decorated with biologically active molecules, for diagnostic tasks such as the detection and capturing of DNA, proteins, cells, and others from biological samples or even from living organisms is known.

WO 2006131400 teaches the decoration of a stainless steel wire with metallic islands that are modified with antibodies for specific cell capture. The metallic islands with sizes in the 100 nm regime were fabricated by using a sphere-monolayer as a shadow mask during the deposition step of a gold layer. The gold islands were modified with thiolated linker molecules that bind specific antibodies.

In EP1907848 a diagnostic-nanosensor is described, e.g., in the form of a catheter or spring wire, consisting of a carrier comprising areas on two-dimensioned arched metallic nanostructures with detection-molecules. The diagnostic nanosensor may be used for the direct detection and isolation of rare molecules or cells out of the peripheral blood or the body. This application technique enables diagnostic procedures that were not possible before: prenatal diagnoses of chromosomal aberration using fetal trophoblasts present in maternal circulation; cancer diagnoses and monitoring of cancer therapy based on the detection of disseminated cancer cells in the body.

EP2344021 relates to a device for detecting analytes, comprising a polymer fiber and capturing molecules, wherein the capturing molecules bind to an analyte and/or a linker molecule. The capturing molecules are selected from the group comprising antibodies, antigens, receptors, polynucleotides, DNA probes, RNA probes, polypeptides, proteins and/or cells. The functionalized polymeric fiber described in this document as having microstructures or surface geometries on the surface can be introduced into a biological sample, such as a blood sample or into a vein of a living organism. During a period of at least a few seconds and several hours, the fiber collects the respective target analyte through its biofunctional coating. After the collection process has finished, the fiber is retracted and the captured material is separated from the fiber for analysis.

Likewise from EP2547250 a biodetector is known with a functionalized surface for isolating molecules or cells from the human body. This biodetector is introduced into the human body for the isolation and enrichment of target molecules and target cells and after a short period is once again removed from the human body.

In “An ensemble of aptamers and antibodies for multivalent capture of cancer cells”, in Chem. Commun., 2014, 50, 6722, by Jinling Zhang et al, an optimized ensemble may be found that functions as a multivalent adhesive domain for the capture and isolation of cancer cells. The ensemble may be incorporated into a microfluidic device that contains micro-pillar arrays. The antibodies and aptamers were immobilized onto the microfluidic channels using avidin and biotin reactions as described in “Aptamer-enabled Efficient Isolation of Cancer Cells from Whole Blood Using a Microfluidic Device” by Sheng W, Chen T, Katnath R, Xiong XL, Tan WH, Fan ZH. Anal. Chem. 2012;84:4199-4206. The microfluidic device is reported to yield a capture efficiency of >95% with a purity of ~81 % at a flow rate of 600 nL/s. This demonstrates technical feasibility, but the flow rate is insufficient to be clinically applicable. More precisely, at this rate one would need more than 46 hours of continuous operation at 100% efficiency to get a hypothetical yield of 100 CTCs, which is not practical at all.

Tumor cells are currently isolated in vitro with high sensitivity and selectivity for subsequent screening of clinically relevant parameters. However, given the very low occurrence of CTCs in blood, even the largest blood sample that can be taken still gives very little information. Further improving the sensitivity and selectivity will allow instead to do so in vivo, which will greatly help further screening of clinically relevant parameters.

The capture of bioanalytes, such as circulating tumor cells (CTCs), directly from within the bloodstream in a human vein is challenging. As indicated above, capturing bioanalytes is typically done with antibodies, which are highly selective and provide a strong and near permanent bond. However, antibodies are only able to form these bonds at very low relative speeds and they rapidly lose their capability to capture bioanalytes at speeds above 1 mm/s, whereas the speed within the bloodstream is multiple orders of magnitude higher still. Moreover, a problem in the state of art is that trapped CTCs, are often damaged or destroyed while retracting the device after capture or upon release. As a result, detection or diagnosis is no longer possible or trustworthy.

Accordingly, there remains a need for a medical sampling device with improved sensitivity and selectivity to isolate CTC’s from a circulating system, preferably from the circulatory system, more preferably directly in the bloodstream, i.e. a modified guidewire or catheter with increased capture rate that allows for release for further screening of clinically relevant parameters.

In a paper by Zhi Sheng-liang et al, “Fabrication of Carbohydrate Microarrays on Gold surfaces: Direct Attachment of Nonderivatized Oligosaccharides to Hydrazide-modified Self- Assembled Monolayers”, Analytical chemistry, part 78, No. 14, 2006, pp. 4788-4793, a process is disclosed for applying a coating comprising heparin on the surface of a microarray, wherein the heparin is end-point attached to amine groups on its surface. The microarray can be used for mapping carbohydrate-protein recognition events.

W02010019189 describes inter alia a medical device having a surface which comprises a coating layer based on end-point attached heparin, which is covalently attached to said surface through a link comprising a 1 ,2,3-triazole. The heparin is used as an anti-coagulant compound. The medical device is not used as a sampling device. Although branching is mentioned this is neither of the heparin, nor is it mentioned to improve the capture rate.

In WO2013188073, a method is described for the removal of mediators that contribute to the pathogenesis of cancer from blood by contacting the blood outside of the body with a solid. This is therefore not a diagnostic device. The method is specifically described for heparin as other carbohydrate surfaces may be significantly less blood compatible than heparinized surfaces and can lead to increased thrombogenicity.

The main application of this invention is to improve upon existing methods to isolate CTCs by providing a coating that functions as an additional selection mechanism and increasing the sensitivity of existing selection mechanisms while also imparting hemocompatibility. The coating can be applied in in-vivo enrichment tools such as modified guidewires and catheters that capture CTCs directly in the bloodstream (e.g. Gilupi CellCollector®). However, the coating may also be applied in in-vitro technologies such as magnetic beads separation (e.g., The CELLSEARCH® Circulating Tumor Cell Kit, which is intended for the enumeration of CTCs of epithelial origin (CD45-, EpCAM+, and cytokeratins 8, 18+, and/or 19+) in whole blood) and some microfluidic flow cells as well. of the Invention

The present invention provides a coating comprising hyaluronic acid (HA), for application on the surface of a medical sampling device, wherein the HA is end-point attached to the surface of the medical device. Moreover, it provides a process for applying a coating comprising HA on the surface of a medical sampling device, wherein the HA is end-point attached to the surface of the medical device, comprising a step (A) of aminating the surface, followed by a step (B) of end-point attaching the HA to the amine groups on the surface, optionally followed by a step (C) of blocking any residual amine groups on the surface. The present invention also provides the medical sampling device for capture of circulating tumor cells having the coating applied thereon.

The present invention also provides a method for the capture of circulating tumor cells and a method for release and analysis of the captured circulating tumor cells.

Fig. 1, A-D is a schematic presentation of the amination and application of the prior art coating on a substrate applied on the surface of a medical sampling device, based on polysaccharides that are not end-group attached.

Fig. 2, E-G is a schematic presentation of the application of the present coating on the substrate, referred to as direct brush, based on end-point attached hyaluronic acid, with the active groups on the surface of the substrate being blocked.

Fig. 3, H-M, is a schematic presentation of the application of the present coating on the substrate, referred to as indirect brush, with the hyaluronic acid being end-functionalized and attached through a linker.

Fig. 4, R-S, is a schematic presentation of the application of the present coating on the substrate, referred to as spaghetti brush, with the linker being a polysaccharide, preferably HA, but now reacted via the side-groups.

Fig. 5, N-Q, is a schematic presentation of the application of the present coating on the substrate, referred to as bottle brush, with an additional HA end-group attached on the HA that is end-group attached to the surface.

Fig. 6, T-W, is a schematic presentation of the application of the present coating on the substrate, wherein receptors and/or ligands are applied onto the hyaluronic acid

Fig. 7, (a)-(d), is a series of histograms illustrating the effect of the coating on the cell velocity and on the capture rate. Detailed description of the invention

The medical sampling device according to the present invention has an improved sensitivity and selectivity to isolate CTC’s directly in the bloodstream. It may be in the form of a guidewire or catheter. Importantly, it shows an increased capture rate whilst at the same time allowing for release of the captured CTC’s for further screening of clinically relevant parameters.

The process of the present invention starts with the amination of the surface of the medical sampling device.

Step (A) of the present process comprises the amination of the surface of the medical device. Preferably, the amination is performed with a diamine.

Step (B) of the present process comprises the end-point attachment of HA to the amine group. HA may be directly attached, which herein is referred to as direct brush, or through a compound acting as a linker. The linker may be a simple bifunctional compound, which is herein referred to as the indirect brush, or may be a polysaccharide itself, but now connected to the surface by some of its side groups, with the remainder of the side groups being available for end-point attaching HA.

In direct brush, the hyaluronic acid is end-point attached to the surface preferably by way of reductive amination. The reductive amination may be carried out in the presence of a reducing agent. It is preferably carried out in the presence of sodium cyanoborohydride.

In indirect brush, the process comprises the end-point attachment of HA to the linker that has reacted with the amine group. Preferably hyaluronic acid is used, onto which a diamine is grafted by way of reductive amination to its terminal end. The reductive amination may be carried out in the presence of a reducing agent. It is preferably carried out with adipic dihydrazide in the presence of sodium cyanoborohydride.

In spaghetti brush, a polysaccharide, which may be HA is reacted with the amine groups on the surface. Different from the direct brush process in this case the side-groups of the HA reacted with the amine groups on the surface, resulting in the orientation of the polymer molecules parallel to the surface, hence “spaghetti”. End-point attachment of HA can then be arranged by reacting for instance with HA onto which a diamine is grafted by way of reductive amination to its terminal end. The end-point attachment is therefore similar to the process of indirect brush. To further enhance the efficiency of the coating, it is preferred that the process comprises an additional step of end-point attaching HA onto the already end-point attached HA. This may be done by carbodiimide coupling, although other reactions are possible. This process may be repeated twice or more often, to create layers of HA grafted on HA, grafted on HA. The HA may be identical, but may also differ in molecular weight, or have a different degree or type of substitution (discussed hereafter).

End-point attachment of HA, however, leaves unreacted amino groups on the surface of the medical sampling device. These residual amino groups will negatively affect the selectivity of the applied coating and therefore need to be blocked. Preferably, these residual amino groups are blocked in step (C) under mild conditions, without affecting the end-point attached HA. For instance, residual amino groups may be converted into aldehyde groups, and these residual aldehyde groups may be blocked by reaction with an amino acid, thereby creating a free acidic group that no longer affects the selectivity of the applied coating. For stability reasons, the imine bonds may be reduced, preferably with the use of sodium cyanoborohydride.

In the present invention, HA is used, which shows a surprisingly high binding affinity for bioanalytes such as CTCs.

By using end-point attachment, and due to their negatively charged side groups, the HA molecules repel each other and may therefore project away from the surface of the medical sampling device, extending farther into the blood, thereby maximizing its availability for binding specific receptors on tumor cells. Typically, the coating has a thickness in the range of 0.1 m to 2pm.

Accordingly, the coating of the present invention comprises or even consists of hyaluronic acid (HA), a hemocompatible glycosaminoglycan that is native to the human body. HA is preferred for several reasons, as discussed hereafter.

Due to its carboxylic acid side group, HA can be easily chemically modified (hereafter substituted HA) and coated onto surfaces. Moreover, the side groups of HA allow for facile coupling of additional molecules such as antibodies and other receptors.

HA is of particular interest, as it has been discovered that tumor cells often show abundant expression of hyaladherins such as CD44, receptors that specifically bind HA. Most tumor cells can therefore adhere to a coating of HA, whereas HA is otherwise generally repellant and non-fouling. By using end-point attachment, HA is applied such that it is maximally available for binding specific receptors on certain types of cells, most notably tumor cells.

The interaction between CD44 and HA is not limited to simple cell adhesion. The present inventors have demonstrated that CD44-positive cells can roll over HA coated substrates. This interaction might play a role in the extravasation and homing of immune cells, and might therefore also be involved in CTC extravasation and metastasis. The ability of HA-coated surfaces to induce cell rolling has been found to be very beneficial in the enrichment of CTCs, because the rolling action likely reduces the speed. This is beneficial because; the higher the flow velocity of a CTC along the surface of a coating, the lower the chance of binding the cell to the coating. In other words, besides the ability of tumor cells to specifically adhere to HA- coated surfaces, tumor cells in flow may also be more easily captured due to its rolling action.

Conveniently, HA can be produced by bacterial fermentation, which avoids the potential toxins and pathogens of animal-derived HA. Bacterial fermentation has enabled industrial production of HA, as evidenced by numerous clinical and cosmetic products, and even dietary supplements. Preferably, the HA has a molecular weight in the range of 40 kDa to 2 MDa, preferably in the range from 50 kDa to 1.5 MDa. The HA may also be substituted, wherein at least some of the functional groups along the polymer backbone have been substituted with other functional groups.

Importantly, HA can be specifically degraded by enzymes under mild conditions thereby allowing controlled release of the CTCs. For instance, HA can be selectively degraded under mild conditions using hyaluronidases, which has also seen clinical applications. In summary, HA is a versatile molecule with many applications in cancer therapy and diagnosis. As a coating, it can provide a unique combination of selective tumor cell adhesion, hemocompatibility and rich chemistry. In other words, the HA coating is optimized for the capture of circulating tumor cells from whole blood, both in-vivo and in-vitro.

As mentioned above, the binding affinity of HA may be enhanced by adding receptors and/or ligands for CTC’s thereon, such as antibodies, preferably monoclonal antibodies, chimeric antibodies, humanized antibodies, antibody fragments or amino acid structures and amino acid sequences, nucleic acid structures or nucleic acid sequences, and the like. For instance, the interaction of HA and aptamers and antibodies is such that the end-point attached HA, the aptamers and the antibodies together are believed to slow down the passing CTC’s and therefore increase the selective capture. As mentioned above, the coating of the present invention is preferably provided by first introducing an amine group on the surface of the sampling device and then end-point attaching the HA. Amination can be achieved best on a medical sampling device having a polymeric surface. This can be accomplished, for example, by the aminolysis of ester- containing polymers such as polyurethanes (Pll), polyesters (e.g. PET) or polymers containing esters in their side group (e.g. PMMA). In principle any polymer capable of aminolysis may be used. Surface amination can also be achieved through silanization, or ammonia-based plasma treatment. Aminolysis is preferably carried out with a diamine, more preferably with ethylenediamine or hexamethylene diamine.

Next, in direct brush, HA may be end-point attached through direct reductive amination. This strategy combines the reductive amination of HA with the coupling of HA to the substrate in a single reaction. The main advantage of this strategy is its simplicity and limited by products, as typically only few steps are required with minimal chemical components. For instance, HA may be directly linked to an aminated surface via surface amines. The reductive amination may be performed using HA, borate buffer ingredients and reducing agents (e.g., sodium cyanoborohydride currently, or sodium triacetyloxyborohydride).

The previous steps have resulted in a substrate with the HA end-attached to the surface. However, the end-point attachment reaction is unlikely to couple a HA molecule to every available surface amine. Consequently, there will be residual surface amines, which is disadvantageous for two reasons: the residual surface amines are a moiety that may provide non-specific adhesion sites, and the positive charge of residual surface amines at physiological pH may cause the HA that is end-point attached to flatten on the substrate. Hence, the residual surface amines must be blocked with a functional group that prevents interactions between the substrate and the HA. The blocking of residual surface functional groups can also prevent undesirable chemical reactions between the substrate and other reactants in downstream reactions, such as the coupling of detection receptors. As mentioned above, this is preferably achieved with a dialdehyde and subsequently an amino acid, because these reactions are simple, quick and do not affect the end-point attached HA. The dialdehyde can be any molecule with two or more aldehyde groups, but glutaraldehyde is preferred because it is readily available and reacts efficiently. The dialdehyde will decorate the substrate with unreacted aldehydes, which can then be reacted with any amino acid based on the desired properties. Preferably 6-aminocaproic acid is used, because it does not have any side groups. Other good alternatives would be aspartic acid and glutamic acid, as these amino acids contain an additional carboxy moiety, which is advantageous for attaching receptors (discussed hereafter). The medical sampling device according to the present invention may therefore be made by applying first a polymeric substrate thereon, which is then aminated, followed by the end-point attachment of HA, whereupon preferably any residual amine groups on the surface are blocked. Optionally the HA, is then further modified, step (E), to add further detection receptors thereon.

For instance, the carboxy groups of HA are easily decorated with detection receptors if the detection receptors have amino groups, which is commonly the case with antibodies and other proteins. Aptamers, if necessary, can be functionalized with an amino group. Decorated HA with detection receptors with available amino groups is preferable because simple and effective carbodiimide coupling can be used. Other non-carbodiimide coupling methods may be used to decorate HA with detection receptors, either with or without intermediate linkers and depending on the moieties available on the detection receptor.

The legend for Fig.1-6 is as follows:

X01 polymeric substrate

X02 amino group on substrate surface

X03 polysaccharide (HA)

X04 polysaccharide side group

X05 activated polysaccharide side group

X06 polysaccharide side group coupled to amino group of substrate surface

X07 polysaccharide reducing terminal end coupled to amino group of substrate surface

X08 diamine

X09 reducing terminal end of polysaccharide

X10 resulting amino group at reducing terminal end of polysaccharide

X11 aldehyde on substrate surface

X12 amino group of reducing terminal end of polysaccharide coupled to surface aldehyde X13 inert chemical group on substrate surface (carboxy group in example)

X14 amino group of reducing terminal end of polysaccharide coupled to carboxy side group of end- attached polysaccharide

X15 polysaccharide end-attached to end-attached polysaccharide

X16 polysaccharide end-attached to side group of attached end attached polysaccharide X17 receptor (antibody) attached to side group of carbohydrate

In Fig. 1 (A-D) an artist impression is provided of a coating wherein a polysaccharide X03 is applied onto a polymeric substrate X01 without end-group attachment. In this prior art process, amine groups X02 are provided onto the substrate. The polysaccharide side-groups X04 are activated, creating X05 which react with the amine groups X02, forming connections X06 whereby the polysaccharide is coupled to the amino group of the substrate surface. This is colloquially referred to as “spaghetti”.

Turning to Fig. 2, a series of steps are provided to apply the coating of the present invention as “direct brush” onto the polymeric substrate X01. In Fig. 2E, HA is end attached to the substrate forming a connection X07. In Fig. 2F, the residual amino groups are blocked with dialdehydes, forming aldehydes groups X11 on the surface, whereas in Fig. 2G the aldehyde groups are blocked to form a chemically inert group X13 on the substrate surface, e.g., with an aminoacid.

Turning to Fig. 3, a series of steps are provided to apply the coating of the present invention as “indirect brush” onto the polymeric substrate X01. In Fig. 3H-I, the HA X03 is reacted at its end X09 with a diamine X08, preferably dihydrazide, resulting in an end-group X10 that facilitates the end-group attachment. In Fig. 3J the substrate is reacted with a dialdehyde to form an aldehyde group X11 on its surface. In Fig. 3K, the end-group X10 is connected with the aldehyde group X11, forming the connection X12. In Fig. 3L residual amino groups are blocked with dialdehydes, whereas in Fig. 3M the aldehyde groups are blocked with aminoacids.

Turning to Fig. 4, a series of steps are provided to apply the coating of the present invention as “spaghetti brush” onto the polymeric substrate X01. In Fig. 4R , the HA X03 is reacted at its side groups X05, as described in Fig. 1. In Fig. 4S, the end-group functionalized HA X16 is connected with a free functional group of the spaghetti HA X05, forming the connection X12.

Turning to Fig. 5, a series of steps are provided to apply the coating of the present invention onto the polymeric substrate X01. In Fig. 5N, the side groups of the end-point attached HA are activated, X05, for subsequent coupling with X10, resulting in bond X14 in Fig. 80, through which HA X15 is end-attached to X03. In Fig. 5P residual amino groups are blocked with dialdehydes, whereas in Fig. 5Q the aldehyde groups are blocked with aminoacids. Finally, by repeating the steps in Fig. 5N-0 a HA X16 is end-attached to X15 in Fig. 5Q.

In Fig. 6, T-W, receptors X17 are provided onto side-groups X04.

Turning to Fig. 7, (a)-(c) histograms are shown that illustrates the effect of the grafting method on the velocity of cultured breast cancer cells in a flow cell. Each histogram contains a positive and negative control, with 'spaghetti HA' as baseline. It should be noted that in absence of any interaction between the coating and the cells (- control), the vast majority of cells are flowing too fast to be quantified. Vice versa, cells that interact so strongly with the coating that they adhere, will be counted as the lowest velocity bar (<5 pm/s). Consequently, the higher this bar (at 5 pm/s), the more a coating can capture cells. A detailed discussion is provided in the experimental section below. In bottlebrush 2 and bottlebrush 3 the grafting step of the experimental section is repeated once or twice respectively. Repeating the grafting step results in a higher number of branching points and amount of grafted HA chains.

The effect of HA on slowing down and adhering cells is illustrated by the following experiments, in which test samples in the form of a coated substrate are tested in a flow cell. The efficiency is compared with two controls: with a COOH coating and an NH2 coating. It is also compared with a coated substrate where the HA is neither end-point attached nor extending into the flow.

Examples

Materials used: 1. Surface amination

Aminolysis is a reaction where an amine reacts with an ester to form an amide bond. EDA was used because it reacts efficiently. The protocol described below was used. a. thoroughly clean the substrate by hand with MQ and detergent b. ultrasonicate the substrate at 40kHz in MQ with detergent for 60 minutes at room temperature with the use of a Branson 5510 ultrasonicator c. wash the substrate with MQ d. ultrasonicate the substrate at 40kHz in MQ for 60 minutes at room temperature e. let the substrate dry f. prepare 2M EDA in 96% EtOH g. incubate substrate in EDA solution for 2 hours (hrs) h. gently rinse the substrate with 96% EtOH i. wash the substrate in 96% EtOH for 2+ hrs under light stirring j. repeat the previous step with fresh 96% EtOH k. dry for 2+ hrs at 40°C.

An aminated substrate will function as a positive control in the flow setup experiment, as amino groups are a moiety well known for their ability to non-specifically adhere many cells.

2. Reductive amination

The goal of this step is to end-point attach the HA with a reducing terminal end to the animated substrate. This is achieved by directly coupling the surface amine to the aldehyde (which is in equilibrium with a lactol) on the reducing terminal end of the HA, and reducing the resulting bond with a reducing agent. Sodium cyanoborohydride was used as the reducing agent, because it is relatively mild and readily available. a. dissolve 0.1M Na2Br4O? and 0.4M NaCI in MQ b. adjust the pH to 8.4 with HCI (using a VQS-70005 pH meter) c. bubble the resulting buffer with N2 gas for 2+ hrs d. dissolve 5% w/v HA and 0.07M NaBHsCN to the volume of borate buffer e. incubate the substrate in the prepared borate buffer at 40°C for 5 days f. gently agitate frequently during incubation g. gently wash the substrate with 1X PBS pH=7.4 three times or more

A substrate treated without reductive amination will serve as a negative control in the flow cell experiment. This substrate will be ultimately decorated with carboxyl groups only, which is a moiety well known for its limited non-specific cellular adhesion. 3. blocking of residual surface groups

The previous steps have resulted in a polyurethane substrate with hyaluronic acid end- attached to the surface. The reductive amination reaction is unlikely to couple a hyaluronic acid molecule to every available surface amine. Consequently, there will be residual surface amines, which is disadvantageous for two reasons: the residual surface amines are a moiety that may provide non-specific adhesion sites, and the positive charge of residual surface amines at physiological pH may cause negatively charged hyaluronic acid that is end-point attached to flatten on the substrate.

The residual surface amines are blocked by reaction with a dialdehyde (GA) and subsequently an amino acid (6AC), because these reactions are simple, quick and do not affect the end-point attached hyaluronic acid. a. prepare 2% v/v GA in MQ b. incubate the substrate in the GA solution at RT for 6+ hrs c. gently agitate frequently d. gently wash the substrate with MQ three times or more e. prepare 100mM 6AC in MQ f. incubate the substrate in the 6AC solution for 6+ hrs g. gently agitate frequently h. gently wash the substrate with 1X PBS pH=7.4 three times or more

4. (optional) decorating of the HA with detection receptors

The protocol below describes one method for carbodiimide coupling with the anti-EpCAM antibody VLI1D9. There are also other anti-EpCAM antibodies available, but the protocol will most likely be the same for most (if not all) IgG 1 antibodies. a. dissolve 100mM NHS and subsequently 100mM EDC in 0.1M MES buffer pH=5.5 b. incubate the polymer surface with the EDC/NHS solution at RT for 2 hrs c. gently agitate frequently d. gently wash the substrate with 1X PBS pH=7.4 e. prepare a 5 pg/ml solution of VI11 D9 f. incubate the polymer surface with the VU1 D9 solution at RT for 2 hrs g. gently agitate frequently h. gently wash the substrate with 1X PBS pH=7.4

5. spaghetti HA preparation (Comparison or intermediate for spaghetti brush) The previous steps have produced a sample with end-point attached HA. To compare the effect of the end-point attached HA to spaghetti HA, the aminated sample of Pll is treated as follows: a. dissolve 5% w/v 50 kDA HA in 0.1 M MES buffer pH=5.5 b. add 100mM NHS and subsequently 100mM EDC gently agitate frequently c. incubate the polymer surface with the HA solution at RT for 3+ hrs d. gently agitate frequently e. gently wash the substrate with 1X PBS pH=7.4 three times or more.

6. Indirect Brush

Instead of step 2, this process involves adding a linker to the aminated polymer surface, and end-point attaching HA. This may be achieved by coupling a dialdehyde to the surface aminogroups, providing a reactive aldehyde on the surface that can subsequently bind to the amino group of the end-aminated HA. Most dialdehydes should work, but we prefer relatively short and simple dialdehydes like gluteraldehyde because they limit the chance of both aldehydes coupling to the surface. a. dissolve 2% v/v GA in MQ b. incubate the polymer surface in the GA solution at room temperature (RT) for 6+ hrs c. gently agitate frequently d. gently wash the polymer surface with MQ three times or more.

Next step is to couple the end-aminated HA to the aldehyde-functionalized surface, thereby creating end-point attached HA. The reaction of the amino-group on the reducing terminal end of the end-aminated HA will react spontaneously to the surface aldehyde. a. dissolve 1 % w/v of end-aminated 1.5 MDa HA in MQ b. incubate the polymer surface with the iHA solution for 6+ hrs c. gently agitate frequently d. gently wash the polymer surface with 1X PBS pH=7.4 three times or more.

Typically, this is followed with steps 3 and 4.

7. Spaghetti Brush

Instead of step 2, this process involves adding a polysaccharide to the aminated polymer surface, as described in step 5, and attaching end-aminated HA, similar to the coupling reaction of step 6. Typically, this is followed with steps 3 and 4.

8. Bottle Brush Subsequent to step 3, HA is grafted onto the end-point attached HA. Grafting may be carried out by the following method: a. dissolve 100mM NHS and 100mM EDC in 0.1M MES buffer pH=5.5 b. incubate the polymer surface with the EDC solution for 1 hrs at RT c. gently agitate frequently d. gently wash the substrate with 1X PBS pH=7.4 e. dissolve 1% w/v iHA in 1X PBS pH=7.4 f. incubate the polymer surface with the iHA solution at RT for 5+ hrs g. gently agitate frequently h. gently wash the polymer surface with 1X PBS pH=7.4 three times or more.

Typically, this is followed with step 4.

9. Cell suspension preparation

In order to test the effect of end-point attached hyaluronic acid on the adhesion and rolling of cancer cells in flow, cancer cells have been cultured and flowed along the treated substrate. The MCF7 cell line was used for their robust nature and expression of CD44 and EpCAM. The first step is to prepare a suspension of cells. a. culture the MCF7 cells to confluency according to the protocol provided by ATCC b. harvest MCF7 cells with Accutase® according to the protocol provided by Innovative Cell Technologies c. stain MCF7 cells with CM FDA according to the protocol by ThermoFisher d. suspend the MCF7 cells in DMEM+10% FBS at a concentration of 20.000 cells/ml

10. Flow cell

To test the effect of end-point attached hyaluronic acid on the adhesion and rolling of cancer cells in flow, a setup was used that consists of a modified syringe pump, a custom flow cell assembly and epifluorescence microscope. The syringe pump was modified in order to pump the cell suspension through tubes connecting to a channel in the flow cell, where an average velocity of 1mm/s was maintained, while the epifluorescence microscope captures images of the luminal surface of the treated substrate for 20 minutes.

The requirements of the setup are as followed: a. a syringe pump: i. with two 5-60ml syringes ii. with vertically mounted syringes iii. with individually configurable infusion/withdrawal rate for syringes iv. with controlled heating to 37°C for both syringes v. with configurable infusion/withdrawal rates between 0.001 -128ml/min b. a flow cell assembly: i. capable of providing an even clamping pressure of 3.2 MPa on an assembly of

1. a microscopy-slide sized treated substrate 1 mm thickness

2. a microscopy-slide sized PDMS gasket of 0.5mm thickness

3. an Ibidi sticky-Slide I Luer with 0.6mm channel height (80188) ii. with controlled heating to 37°C of the treated substrate c. an epifluorescence microscope: iii. with a 420-490nm bandpass excitation filter iv. with a 520nm longpass emission filter v. with a 4X objective vi. with a camera capable of capturing 1280*960 JPEG images with RGB24 format at 100ms exposure time at ISG100.

11. Data analysis

The images as captured in the previous step are analyzed to quantify the effect of the endpoint attached hyaluronic acid on the flow- and rolling velocity and immobilization rate. The data analysis is performed as follows: a. the images are preprocessed in Fiji (an open source image processing package based on Imaged) with a custom macro whereby i. the green channel is isolated ii. the 'Subtract Background" function with rolling ball radius of 50 is used iii. a static value is subtracted from all images b. the images are analyzed with the Trackmate plugin iv. DoG detector with 20px diameter and 1.0 threshold v. LAP tracker with 30pix distance and 50 Y penalty c. the tracks are exported and processed in MS Excel vi. all tracks with <5 spots are deleted vii. a histogram with 0.25 bin size is made viii. the bin sizes are converted to pm/s d. the histograms are compared to a positive and negative control e. the histograms are compared to non end-point attached HA The histograms in Fig. 7 show the distribution of travel velocities of cells/cel I clusters over various substrates. All cells/cel I clusters that could not be tracked for at least 5 frames were discarded. Each bar represents a range of velocities, i.e., 0-4.9, 5-9.9, 10-14.9, 15-19.9, 20- 24.9, and 25-30 pm/s. Cells/cell clusters that were completely immobilized are included in the 0-4.9 pm/s bin. Cells/cell clusters that were traveling faster than 30 pm/s are not included.

The cell suspensions were equal in concentration, meaning that if cell/cell cluster counts are low, the cells generally did not interact with the substrate or adhere to it.

The negative control sample is decorated with -COOH groups that do not interact with the cells or cell clusters and prevent attachment to the substrate, as evidenced by the very low 0- 4.9 pm/s bin. The positive control sample is decorated with -NH2 groups, which are known to adhere cells well. The spaghetti sample is coated with HA that is coupled to the substrate along the length of the molecule, which results in a marginal increase in the rolling and adherent cells/cell clusters. The sample with end-point attached HA shows many more cells/cell clusters that adhere or interact. Because spaghetti HA and end-point attached HA differ only in the method of attachment, the data shows that end-point attachment greatly improves the interaction and adhesion of cells/cell clusters with the substrate.

12. Releasing CTC’s

Captured CTCs bound to the coated substrate can be gently released from the substrate through enzymatic degradation of the HA. Unlike classic approaches like trypsinization, this approach minimally affects the viability and phenotype of the CTCs and is therefore ideal for a subsequent analysis that requires unaffected cells. Hyaluronidase from bovine testes is preferred, because it is selective, efficient and economical. Other enzymes that can degrade HA or hyaluronidases from other sources may be used instead. The protocol below describes a simple incubation with a hyaluronidase solution. a. prepare a solution of 200 U/ml hyaluronidase in 1X PBS pH=7.4 b. prewarm the solution to 37°C c. incubate the substrate with CTCs at 37°C for 5min d. harvest the solution to obtain the CTCs in suspension

Viable CTC’s with unaffected phenotype can be obtained, which is very difficult to achieve and which is highly desirable for further for further screening of clinically relevant parameters.