WO2023036728A1 - Microfluidic platform for interactions studies - Google Patents

Abstract

There is provided a microfluidic system comprising a) a microfluidic device and, b) a plurality of polymer microgels, wherein the microfluidic device comprises at least one fluid inlet (1,2), at least one fluid outlet (6), at least one main channel (5), in fluid contact with the fluid inlet (1,2) and the fluid outlet (6), a plurality of hydrodynamic traps in fluid contact with the at least one main channel (5), wherein each hydrodynamic trap is an open cavity with a size and a shape which are adapted to trap and accommodate one microgel, wherein each hydrodynamic trap has an inlet and at least one outlet, wherein the outlet is downstream of the inlet in the at least one main channel (5) in an intended direction of flow in the main channel (5). Further a method utilizing the system is provided.

Description

Microfluidic platform for interactions studies

Field of the invention

The present invention relates to a system and a method for studying interactions between at least one compound and polymers in microgels .

Background

Parenteral delivery is today the most common approach in formulation of pharmaceutical peptides and proteins . However, while avoiding the maj or di f ficulties associated with oral uptake , such as degradation in the digestive tract and first-pass metabolism, the fate of subcutaneously administered drugs is often di f ficult to predict . For many drug products , the fraction of drug molecules reaching the circulatory system is low and there is often large variation between patients which is a problem shared with oral delivery of macromolecules . Aggregation of the drug after inj ection and interaction with the constituents of the extracellular matrix (ECM) in the subcutaneous environment are factors believed to contribute to the limited bioavailability and poor absorption predictability in addition to catabolic degradation . From this perspective , it is important to be able to investigate , early in the drug development phase , how drug candidates interact with the constituents of the ECM . In the present work, we describe and evaluate a microfluidics method for probing interactions based on the responsiveness of polyelectrolyte microgels to oppositely charged species . As responsive microgels are used as drug carriers in, e . g . , transarterial chemoemboli zation ( TACE ) , the method is suitable also for in vi tro studies of microgel-based delivery systems . Collagen, hyaluronic acid and chondroitin sul fate are the maj or macromolecular species of the extracellular matrix in subcutaneous tissue . Collagen triple helices form a weakly charged, three-dimensional network, largely responsible for the mechanical rigidity of the tissue . The pores in the network are large enough not to sterically hinder the transport of small peptides , but can af fect the mass transport of larger species like , e . g . , antibodies . Hyaluronic acid and chondroitin sul fate are negatively charged glycosaminoglycans dissolved in the interstitial fluid between the collagen fibres . The highly charged chondroitin sul fate ( ~ 1 negative charge per 5 A along the chain) is expected to interact electrostatically with net positively charged peptides , thereby af fecting their transport rate . Hyaluronic acid has lower linear charge density ( ~ 1 negative charge per 10 A) but is present at higher concentration, and may therefore af fect the transport properties to an even larger extent than chondroitin sul fate .

The interaction between proteins and hyaluronic acid and other biopolyelectrolytes has been studied extensively in the past . The results from a number of in vitro studies of mixtures of proteins and polyelectrolytes of opposite charge have revealed the importance of electrostatic interactions . Thus , complex formation and phase separation are generally favored by high net charge of both species , low ionic strength, and high flexibility of the polyelectrolyte . However, the interaction patterns also depend on the distribution of charges on the protein and the propensity of both the protein and polyelectrolyte to regulate their net charge . The latter properties have been proposed to explain why complex formation and even phase separation have been observed in mixtures of species of the same charge sign, sometimes described as complexation on the "wrong side" of the isoelectric point . Many results have been obtained from phase studies , by means of turbidimetric titration or analysis of the composition of co-existing phases in phase separated samples , in combination with rheological and structural investigations .

Scissor® is an in vi tro method developed as a tool for predicting the dissolution and absorption of drugs after subcutaneous administration . The drug formulation of interest is inj ected into a solution of high molecular weight Hyaluronic acid, Collagen, Chondroitin sul fate and Fibronectin contained in a cartridge , where di f fusion and aggregation of inj ected formulation can be monitored . Hyposkin® is an ex vi vo method to predict adsorption through the subcutaneous tissue , where full skin tissue samples from abdominal surgery are preserved and placed in a biological matrix to keep the sample alive for at least 7 days . A drug formulation is inj ected into the subcutaneous layer of the tissue sample and concentration measurements are made in a release media below the skin tissue . There are also several in vi tro methods utili zing agarose or Sephadex® hydrogels aiming to mimic the subcutaneous tissue , these are primarily mimicking the pore structure found in the ECM and not the physicochemical interactions known to occur .

Scissor and the other methods j ust described, typically require large amounts of material , not always available in investigations of drug candidates in the development phase . While the existing micropipette/micromanipulator approach for microgels is very useful for detailed investigations of particular systems , it is quite cumbersome to use , and permits only the examination of one microgel at a time . Ahnfelt et al in Journal of Controlled Release 292 ( 2018 ) 235-247 studies single beads as a clinical drug delivery system . The repeatability of the studies of single microgels is a problem and the technique requires great operator skills and long training of the operator to be reliable .

WO2021061522 Al discloses a fluidic device for trapping tissue samples comprising an array of traps shaped to trap a tissue sample . The device is used to predict drug responses in human by addressing portions of a tissue sample with, for example , drug candidate and imaging reagents .

US20210197196 Al discloses a microfluidic chip for trapping a plurality of obj ects comprising trapping channel including a hydrodynamic trap . The obj ect may for example comprise cells , organoids or microspheres . The chip is suited for loading and culturing of cell clusters and organoids in order to perform experiments on them . The chip may be used for identi fying useful compounds by determining their ef fect on cells .

Subcutaneous inj ections are one of the parenteral administration routes commonly used, but there are di f ficulties in predicting the behavior of inj ected biopharmaceuticals in the subcutaneous tissue and with that also the bioavailability which can vary from 10- 80% depending on drug . There is an expressed desire from the pharmaceutical industry for someone to develop in vi tro methods that increase the understanding of the behavior of drug molecules in the subcutaneous tissue and may predict bioavailability .

Summary

One object of the present invention is to obviate at least some of the disadvantages in the prior art and provide a microfluidic system as well as a method for studying interactions between compound (s) and polymer (s) in microgels .

In a first aspect there is provided a microfluidic system comprising a. a microfluidic device and, b. a plurality of polymer microgels, wherein the microfluidic device comprises:

- at least one fluid inlet (1,2) ,

- at least one fluid outlet (6) ,

- at least one main channel (5) , in fluid contact with the fluid inlet and the fluid outlet,

- a plurality of hydrodynamic traps in fluid contact with the at least one main channel (5) , wherein each hydrodynamic trap is an open cavity with a size and a shape which are adapted to trap and accommodate one microgel, wherein the resistance for a microgel flowing in a liquid in the main channel (5) to enter an empty hydrodynamic trap from the main channel (5) is lower than the resistance for the microgel to continue in the main channel (5) , and wherein the resistance for a microgel to enter an occupied hydrodynamic trap is higher than the resistance for the microgel to continue in the main channel (5) , wherein each hydrodynamic trap has an inlet and at least one outlet , wherein the outlet is downstream of the inlet in the at least one main channel ( 5 ) in an intended direction of flow in the main channel ( 5 ) .

In a second aspect there is provided a method for studying interactions between polymers in microgels and other compounds , the method comprising the steps of : a . providing a microfluidic system as described above , b . adding at least one fluid to the at least one fluid inlet ( 1 , 2 ) , wherein the fluid comprises at least one compound to study, c . measuring changes in the diameter of the microgels upon or after exposure to the compound to study .

Further embodiments of the present invention are defined in the appended dependent claims .

The compound to study may for example be an active pharmaceutical ingredient in a drug, also herein denoted as a drug molecule .

The microgels can be used as a model system for subcutaneous tissue .

A first advantage with the invention is the low amounts of drug substance needed to perform an experiment and the possibility to reuse this amount . The exact amount depends on the drug substance and concentration needed, but in one embodiment a flow rate of 80 pl/min and still having a high accuracy of desired concentration a total of 4 . 8 ml/hour would be used . One hour is according to the experiments usually an abundance of time to reach equilibrium within a system ( drug substance-microgel ) .

A second advantages with the invention is that it is built for automation of experimental procedure , with the only manual steps for every experiment being sample preparation and instrumental setup .

Finally, it is possible to build a fully automated large scale screening setup for interaction by adding accessories to the system . With the setup used in the examples with two containers connected to the microfluidic chip is possible to change one parameter automatically, such as ionic strength, concentration of drug substance or switching between two types of drug substances . By adding a commercially available flow switch matrix it would be possible to automate the flow of at least 15 di f ferent drug substances , with the limitation being that the microgels used need to able to re-swell completely in an appropriate concentration of NaCl without any substantial damage to them .

Kinetics of both binding and release of pharmaceuticals to/ from a microgel can generally be investigated .

Brief description of the drawings Aspects and embodiments will be described with reference to the following drawings in which :

Figure 1 shows hydrodynamic traps with di f ferent flow paths indicated . Path 1 to the left and path 2 to the right . Figure 2 shows a picture of a device used for interaction studies (MIS) taken during the experiment, arrows indicates the direction of flow. (1) Inlet, (2) inlet, (3) mixer (3) for full mixing of different solutions added in the two inlets (1,2) , (4) inlet for microgels, (5) main channel with 96 hydrodynamic traps, each of the 96 traps can hold one microgel. (6) Outlet.

Figure 3 shows swelling of microgel networks as function of the concentration of monovalent salt in the solution (C_sait) • A and B: Impact of network elasticity parameter (p) as indicated. C: Impact of fraction of charged network segments (q) . D: Experimental swelling of microgels (average of 8 microgels) . Swelling is expressed as volume per mole of network charges (V/n_p) or the ratio of the volume (V) relative to the volume in a reference solution containing 5 mM salt (V_o) .

Figure 4 shows volume change of microgels over time when exposed to a single concentration of PRO or CytC in

PBS (5) , flow rate 200 pl/min. Each curve are data for one microgel. A: Polyacrylate (PA) microgel exposed to PRO (10, 250 pM) B: PA microgel exposed to CytC (10,250 pM) , C: hyaluronic acid (HA) microgel exposed to PRO (10,250 pM) , D: HA microgel exposed to CytC (10,50 pM) , E: DC bead exposed to PRO (10, 100 pM) , F: DC bead exposed to CytC (10,250 pM) . Results for one microgel from each experiment presented here.

Figure 5 shows volume ratio (V/V_o) of microgels made out of acrylamide modified hyaluronic acid (denoted HA microgels) at exposure to PRO or CytC at different concentrations in PBS (5) . Flow rate 200 pl/min (80pl/min at 250 pM) . Figure 6 shows volume ratio (V/V_o) of HA microgels at different concentrations of AMT. Flow rate 200 pl/1

(80pl/min for 250 pM) . Data for one microgel.

Figure 7 shows volume change of HA microgels exposed to 22.93 mM AMT in PBS pH 5.9 5 mM. Flow rate 200 pl/min.

Figure 8 shows volume ratio (V/V_o) of HA microgels at exposure to PRO or CytC at different concentrations in PBS (5) . Flow rate 200 pl/min (80pl/min at 250 pM)

Figure 9 shows experimental results of microgel volume response at exposure to 250 pM CytC at different NaCl concentrations .

Figure 10 shows V/n_p (solid curves) and fl (dotted curves) as functions of C_sait for a microgel in equilibrium with a solution of 250 pM oppositely charged spheres (Z=+7; r_s=lnm; p=20, q=l, and C*=2M) . Thick curves ("weak attraction") were obtained by adding an excess free energy of 1.3 per charge for spheres in the microgel.

Figure 11 shows volume change of DC beads measured in MIS over time at exposure to AMT (519 pM) , each curve represent a single different microgel. Total flow rate into microfluidic chip was 200 pl/min.

Figure 12 shows experimental results of volume change at release of AMT from HA microgels at 0 mM NaCl in 5 mM PBS, pH 7.4. Flow rate 200 pl/min.

Figure 13 shows volume ratio of Ha microgels at release of

PRO at different NaCl concentrations 0-1000 mM in PBS (5) . Exposure time 10 minutes for each concentration. Flow rate 200 pl/min (80 pl/min for 1 M NaCl) . Showing results for 8 different microgels.

Figure 14 shows HA microgels exposed to 150 mM NaCl in PBS pH 7.4 5 mM, after binding and release of PRO. Different grades of deformation can be seen.

Figure 15 shows a microscopy image with 4 PA microgels mid-collapse during exposure to PRO 250 pM, in PBS (5) .

Figure 16 shows volume ratio of DC beads and PA microgels at release of PRO at different NaCl concentrations 0-1000 mM for DC beads 0-150 mM for PA microgels in PBS (5) .

Exposure time 10 minutes for each concentration. Flow rate 200 pl/min (80 pl/min at 1000 mM)

Figure 17 shows volume ratio of HA microgel and DC bead at release of CytC at different NaCl concentrations 0-150 mM in PBS pH 7.4 5mM , exposure time 10 minutes each concentration. Flow rate 200 pl/min. First measurement point of HA microgel at 0 mM NaCl (V/V0~ 0.5) is the volume ratio at start of experiment with fully bound CytC, second measurement point at 0 mM NaCl (V/V0>l) is the volume ratio after 10 minutes exposure to only PBS pH 7.4 5 mM .

Figure 18 shows volume change over time of PA microgels at release of CytC at different NaCl concentrations 0-150 mM exposure time 10 minutes each concentration in PBS pH 7.4 5 mM. Exposure time 10 minutes for each concentration. Flow rate 200 pl/min. The figure shows results for 8 different microgels. Figure 19 presents a theoretical model showing the impact of sphere charge on swelling isotherms ; V/n_p is plotted vs . the concentration of spheres ( radius 10 A) with di f ferent charge numbers in a solution with 5 mM salt . Arrows in the figure represents volume phase transitions (VPT ) between swollen and collapsed homogeneous states . The model illustrated how higher charge leads to both larger deswelling and deswelling at lower concentrations of the sphere .

Figure 20 shows a graph showing the determination of critical aggregation concentration of the three di f ferent amphiphilic drugs in polyacrylic acid gels . The drugs being Doxepin, Amitriptyline and Chlorpromazine .

Figure 21 shows a graph showing the collapse of polyacrylic acid microgels at exposure to Amitriptyline , Doxepin or Chlorpromazine at a concentration above the critical aggregation concentration .

Figure 22 shows a graph showing the collapse of Sephadex C-25 ( anionic dextran based microgels ) at exposure to Amitriptyline hydrochloride at a concentration over the critical aggregation concentration ( 20 mM) .

Detailed description

Before the invention is disclosed and described in detail , it is to be understood that this invention is not limited to particular configurations , process steps and materials disclosed herein as such configurations , process steps and materials may vary somewhat . It is also to be understood that the terminology employed herein is used for the purpose of describing particular embodiments only and is not intended to be limiting since the scope of the present invention is limited only by the appended claims and equivalents thereof .

It must be noted that , as used in this speci fication and the appended claims , the singular forms "a" , "an" and "the" include plural referents unless the context clearly dictates otherwise .

The principle behind the technique is that weakly crosslinked polyelectrolyte networks respond to the loading and release of oppositely charged substance species by changing their volume and internal morphology . The responsiveness results from the delicate balance between the elastic forces in the network and the osmotic swelling forces which is sensitive to the concentration of network counterions inside the microgel . The exchange of a few percent of monovalent counterions for a charge equivalent amount of multivalent protein molecules is suf ficient to give rise to a measurable volume change . The ef fect has been studied extensively with polyelectrolyte hydrogels of macroscopic si ze , investigations that have revealed many interesting features about the systems , such as volume phase transitions and phase coexistence in gels . The same phenomena can be studied also by monitoring spherical microgels of diameter 10 - 500 pm using micropipette-assisted microscopy techniques . Maj or advantages with these types of miniaturised experiments is that they consume little material and that the gel response is comparatively fast .

The si ze of polymer coils depends on the quality of the solvent and the interaction with other species present in the solution . Flexible polyelectrolytes e . g . HA, PA and PAMPS are particularly sensitive to the presence of molecules of opposite charge to the polyion chains. When the chains are crosslinked to form three-dimensional networks such as microgels, the interactions affect the osmotic swelling of the network. It is possible to relate the magnitude of the volume response to the nature and strength of the interaction, and therefore to the properties of the interacting species. The method presented in this work is based on that principle.

To make the microgel method more efficient there is an array of hydrodynamic microgel traps on a microfluidic chip. This allows to simultaneously monitor a large number of microgels in contact with the same liquid medium. Thereby it becomes easy to change the medium and the consumption of material is small, making the technique suitable for screening purposes.

In the first aspect there is provided a microfluidic system comprising a. a microfluidic device and, b. a plurality of polymer microgels, wherein the microfluidic device comprises:

- at least one fluid inlet (1,2) ,

- at least one fluid outlet (6) ,

- at least one main channel (5) , in fluid contact with the fluid inlet and the fluid outlet,

- a plurality of hydrodynamic traps in fluid contact with the at least one main channel (5) , wherein each hydrodynamic trap is an open cavity with a size and a shape which are adapted to trap and accommodate one microgel, wherein the resistance for a microgel flowing in a liquid in the main channel (5) to enter an empty hydrodynamic trap from the main channel (5) is lower than the resistance for the microgel to continue in the main channel (5) , and wherein the resistance for a microgel to enter an occupied hydrodynamic trap is higher than the resistance for the microgel to continue in the main channel (5) , wherein each hydrodynamic trap has an inlet and at least one outlet, wherein the outlet is downstream of the inlet in the at least one main channel (5) in an intended direction of flow in the main channel (5) .

A hydrodynamic trap relies on a balance between the resistance of 2 different paths that a microgel may take. Either passing by an empty trap, or entering an empty trap. See Fig. 1. The left path is Pl and the right path is P2. The resistance R, which a microgel experiences when moving is denoted R.

• For a microgel to enter an empty trap, R in P2 < R in Pl

• Once a trap is filled, R in P2 > R Pl, so the microgel should continue its path. I.e., when the trap is filled it is blocked so that another microgel cannot enter.

The resistance R is taken when there is a flow of fluid in the channel and the trap.

A hydrodynamic trap is an open cavity so that a microgel can enter the trap. This open cavity also forms the inlet of the hydrodynamic trap. The size and shape of the trap is adapted to the size and shape of the microgels so that one microgel can fit in a hydrodynamic trap, but so that two microgels do not fit in a hydrodynamic trap. The number of hydrodynamic traps in the microfluidic system is in one embodiment high, as to trap a high number of microgels , in order to get good statistical data from few experiments . In one embodiment , the microfluidic system comprises at least 20 hydrodynamic traps . In one embodiment , the microfluidic system comprises at least 50 hydrodynamic traps . In one embodiment , the microfluidic system comprises at least 80 hydrodynamic traps . In one embodiment , the microfluidic system comprises 96 hydrodynamic traps .

The microgels are arti ficially manufactured responsive polymer networks . Using this system, the interaction between the polymers in the microgel and a compound of interest can be studied . Typically, but not necessarily, the shape of the microgels is spherical and the hydrodynamic traps are adapted to the spherical shape .

The si ze ( diameter ) of the microgels is measured with optical microscopy . The optical microscopy is carried out following the US pharmacopeia 776 Optical microscopy .

Further the methods according to ISO 13322 regarding particle si ze analysis with image analysis methods are followed . Also ISO 9276 ( 2014 ) , "Representation of results of particle si ze analysis" is utili zed .

The diameter of the microgels for the purpose of adapting the si ze of the microgels/ traps is measured when the microgels are as large as possible . The solvent chosen for this si ze measurement is suitably the solvent used during the experiments to be performed using those microgels/ traps for which the microgels will be the largest , so during the experiment itsel f the microgels may vary in size but never go above the size measured when adapting the size of the microgels/ traps . As long as they are not larger than the trap they will stay in the trap. The size of the traps is adapted such that the microgels will at their largest size, measured according to above, fit in the traps without being deformed by the walls of the traps. The walls of the trap should not exert a pressure on the gels such that they are restricted from reaching their largest size, measured when adapting the size of the microgels/ traps when the size of the gels are unrestricted, and the shape of the gels should not be restricted by the size of the traps as they grow. The traps do thereby not render the microgels immobile. The microgels are thus able to move at least slightly in the traps .

For networks without acid/base functionality, the measurement of the size of the microgels is normally conducted in deionized water for the purpose of adapting the size of the microgels and/or traps. Deionized water has a conductivity of about 0.055 microsiemens. The measurement of the size of the microgels for this purpose is carried out starting with deionized water with a conductivity of 0.055 microsiemens or less. Any change of the conductivity of the water during the measurement is disregarded, since that has little influence on the measured results. In one embodiment the method is carried out in a water based solvent and deionized water is used for the size measurement.

For networks with acid/base functionality, the measurement of the size of the microgels for adapting the size of the microgels and/or traps is normally conducted in a buffer with a pH as to obtain a fully charged polymer network, normally resulting in the largest gel volume. The ionic strength of the buffer should be as low as possible. Deionized water typically has no well-defined pH, and may not be the solvent giving the largest gel volume for certain gels with pH-dependent charges.

The environment where the microgels will be the largest is a solvent without unnecessary additives, since additives will generally result in smaller microgels. Additives should be kept to a minimum for the size measurement. If possible deionized water should be used.

For solvents which are not water based the measurement of the size of the microgels is carried out in pure solvent, i.e. solvent without any additions of for instance the compound to study.

The loading of microgels into traps is performed in an appropriate solvent for one trap to fit only one microgel each and avoid double loading.

Thus, the probability for a microgel to enter an empty trap is higher than for the microgel to continue in the main channel. The probability for a microgel to enter an occupied trap so that a trap comprises two microgels is very low, at least when the size of the trap is properly adapted to the size of the microgels

The microgels should preferably have a distribution of their sizes, which is limited and fairly close to a monodisperse size distribution. More in detail this can be expressed so that the average microgel size has a polydispersity index of maximum 1.1 as measured according to ISO 22412:2017. In one embodiment the diameter of the microgels is in the interval 50-500 gm . In one embodiment the diameter of the microgels is in the interval 150-200 gm . In another embodiment the diameter of the microgels is in the interval 100-250 gm, 120- 150 gm, 75- 100 gm, 35-50 gm . In yet another embodiment the diameter of the microgels is in the interval 200-275 gm . In a further embodiment the diameter of the microgels is in the interval 30- 1000 gm . In one embodiment the diameter of the microgels is at least 75% of the width of the traps . It is conceived that the average diameter as calculated from ISO 9276 ( 2014 ) is in the mentioned range . Thus some particles may have a diameter outside the interval although the average is within the interval . For a given average microgel diameter it is necessary to adapt the traps to that particular average microgel diameter . The above intervals are not to be interpreted so that one type of traps can accommodate microgels with an average diameter in the entire range , such as 30- 1000 gm . Instead one average microgel diameter in the interval is to be selected and the traps are to be adapted to that average microgel diameter . A reasonably narrow distribution of the microgel diameters for the individual microgels is preferred . For non-spherical microgels the proj ected area diameter according to US pharmacopeia 776 Optical microscopy is taken to calculate the diameter .

In one embodiment the microgels are crosslinked polymer molecules . They are crosslinked so that they have suf ficient stability during swelling and shrinkage during the experiment . In one embodiment , the microgels comprise at least one polymer, which polymer is naturally present in human subcutaneous tissue . This allows the system to be used as a model system for subcutaneous tissue for studying drug compound interactions in subcutaneous tissue . This without actual human tissue being utili zed .

In one embodiment , the microgels comprise at least one polymer selected from the group consisting of hyaluronic acid, collagen, chondroitin sul fate , heparan sul fate , keratan sul fate and dermatan sul fate . These polymers are common in subcutaneous tissue . Also modi fied polymers which exist naturally in human skin tissue and/or subcutaneous tissue are encompassed . Examples include but are not limited to polymers which have been modi fied to be covalently bound to other chemical groups or polymers .

In one embodiment , the microgels comprise at least one polymer, which polymer is chosen from the group comprising a polymer which is naturally present in human tissue , a synthetic polymer, and a non-human natural polymer . The polymer could be any natural or synthetic polymer which can be formed as a spherical microgel which changes volume as a result from interactions with a compound to be studied could be used . In one embodiment the polymer is a polyacrylate . In one embodiment the microgels are at least one type of microgels selected from the group consisting of DC microgels , pNIPAM based microgels , alignate based microgels , agarose based microgels , dextran based microgels , gelatin based microgels , polyvinylalcohol based microgels , poly ( ethylene glycol diacrylate ) based microgels , poly ( d, 1-lactic acid) ( PLA) based microgels and Polyethyleneglycol ( PEG) based microgels . In one embodiment , the microfluidic system is a closed system which does not allow any or essentially any evaporation of liquid from the system during an experiment .

In one embodiment the microfluidic system comprises a programmable controller communicatively coupled to a camera with an optical system capable of imaging the si ze of the microgels , the controller being adapted to receive an image of the microgels and calculate their volume based on the image , the controller being communicatively coupled to a pump and able to control the flow rate and optionally the concentration of the compound to study, the controller being adapted to calculate and present a result . By using a programmable controller the system can be automated and perform a number of evaluations of recorded data .

In one embodiment each trap has at least one outflow channel , fluidly connecting the trap to a part of the main channel located downstream of the trap, wherein the width of the at least one outflow channel is smaller than the width of the main channel . The width of the at least one outflow channel is preferably smaller than the width of a microgel to be used in its fully collapsed state . A microgel may in its fully collapsed state may have a diameter of for example 80 % of its maximum diameter . Thus the microgel cannot escape from the hydrodynamic trap, at least not through the outlet .

In one embodiment , each hydrodynamic trap has at least two outflow channels . By using such a setup the outlets can be made fairly narrow and the microgels will securely remain in the hydrodynamic traps . In one embodiment, the cross section of the traps have a semi-circular shape. By this shape the hydrodynamic traps are adapted to spherical microgels.

In one embodiment, the main channel (5) comprises a plurality of turns, providing a plurality of channel sections, wherein each section is fluidly connected to a downstream section through at least one trap and the at least one outflow channel of that trap. In one embodiment, the main channel (5) comprises a plurality of turns forming a zigzag pattern with rounded edges, as to avoid clogging of microgels, resulting in a plurality of channel rows, wherein adjacent rows are fluidly connected through at least one trap and the at least one outflow channel of that trap. In one embodiment, adjacent rows run as parallel channels. The hydrodynamic traps are in one embodiment positioned between adjacent rows.

In one embodiment the microfluidic system comprises a mixer (3) , said mixer (3) being positioned in fluid contact with and the at least one main channel (5) , the mixer (3) optionally comprising at least one filter, wherein the mixer comprises at least one flow path wherein the flow path comprises a plurality of features selected from obstacles and turns in the fluid flow path. Such a mixer between the inlet and the main channel (5) ensures proper mixing, especially if there are multiple inlets (1,2) , where different fluids are added. This is useful when varying the concentration of different compounds during an experiment. In one embodiment, the plurality of features is selected from the group consisting of diverting and re-joining channels, flow splitters, sharp edges of the channel, re-occurring width change of the channel, posts and bifurcations. In one embodiment, the microfluidic system comprises at least one element adapted to keep unwanted particles away from the hydrodynamic trap array. Such particles may be dirt or fibers. In one embodiment, the at least one element is located in the mixer (3) . In one embodiment, the at least one element is located in at least one inlet (1, 2) . In one embodiment, the at least one element is a filter .

In one embodiment, the system comprises a separate inlet (4) adapted to add microgels, wherein the separate inlet (4) is before the plurality of hydrodynamic traps and after an optional mixer (3) , wherein before and after are in relation to an intended direction of a flow from the at least one fluid inlet (1, 2) to the at least one fluid outlet (6) . The inlet is adapted to add microgels by having a size so that the microgels can be added directly to the main channel. The advantages is that the microgels can be added to the main channel (5) without passing the mixer (3) , since the mixer (3) may contain a number of obstacles contributing to the mixing, which obstacles may obstruct the flow of microgels in the mixer (3) .

In the second aspect there is provided a method for studying interactions between polymers in microgels and other molecules, the method comprising the steps of: a. providing a microfluidic system as described above, b. adding at least one fluid to the at least one fluid inlet (1,2) , wherein the fluid comprises at least one compound to study, c. measuring changes in the diameter of the microgels upon or after exposure to the compound to study.

The microfluidic system comprises microgels with the polymer of interest. As previously described, the microgels are artificially manufactured responsive polymer networks. The interactions between the compound to study and the polymer (s) in the microgel is studied. The fluid comprising the compound to study passes the main channel and the compound to study interacts with the microgels. The compound to study is preferably dissolved or in the form of a very finely divided suspension. The diameter of the microgels are typically measured by optical measurements of particle sizes according to ISO 13322. For non-spherical microgels the projected area diameter according to US pharmacopeia 776 Optical microscopy is taken to calculate the diameter. The size, and thereby the volume of the microgels changes as a response to interactions with the compound (s) to study. The diameter is measured before and after exposure to the relevant compound and the change is thereby measured. From the change in size, diameter (i.e. volume) , conclusions can be drawn regarding interactions between the polymer and the compound to study and the amount of compound bound to the microgel can be estimated. In one embodiment the diameter is utilized to calculate the volume of the microgels and thereafter relate the volume change of the microgels to the exposure to the compound to study. I.e. the measuring of a change in diameter and thereby volume of the microgels requires at least two measurements in order to establish a change. In one embodiment , the microfluidic system in step a ) is provided by loading a microfluidic device with microgels of a si ze adapted to the si ze of the hydrodynamic traps , by adding the microgels suspended in a fluid to the at least one fluid inlet for microgels ( 4 ) adapted to add microgels . A fluid with suspended prepared microgels is added and due to the hydrodynamic traps the microgels will be placed in the hydrodynamic traps . One microgel in each trap .

In one embodiment of the method, the microgels are loaded during a constant flow of liquid .

The traps and microgels should be experiencing a constant flow of liquid during performance of the method . This minimi zes the stagnant layer .

In one embodiment two liquids are mixed, one being a solvent or a buf fer, and the other being a stock solution comprising the drug compound to be studied to a desired concentration in step b ) .

In one embodiment the concentration of the compound to be studied is changed in a gradient . In one embodiment the composition of the gradient solvent or buf fer or of the solution comprising the compound to be studied is changed in a gradient . In one embodiment the ion concentration in the gradient solvent or buf fer is changed in a gradient .

In one embodiment the concentration of another compound in the gradient solvent or buf fer or of the solution comprising the compound to be studied is changed in a gradient to investigate the ef fect of that other compound on the interactions between the compound to be studied and the microgel . In one embodiment the concentration of the compound to be studied is changed by means of a controller . The controller is in one embodiment communicatively coupled to a pump and able to control the flow rate and optionally the concentration of the compound to study . In one embodiment the composition of the at least one fluid is changed by means of a controller . In one embodiment the ion concentration in the gradient solvent or buf fer is changed by means of a controller . In one embodiment the concentration of another compound in the gradient solvent or buf fer or of the solution comprising the compound to be studied is changed by means of a controller . The concentrations or compositions can thereby be changed automatically without moving the microfluidic system and thereby disturbing the gels in their traps . Disturbing the gels during performance of the method could lead to them leaving the traps and failure of the experiment .

In one embodiment , the interactions between the microgel and the at least one compound to study are observed by measuring the diameter change and calculating the volume change of the microgels . In one embodiment the diameter is measured by light microscopy .

In one embodiment measured data are stored and processed in a programmable computer system .

In one embodiment , the microgels comprise at least one polymer, which polymer is naturally present in human subcutaneous tissue and where the system is utili zed as a model system for subcutaneous tissue . In one embodiment , the microgels comprise at least one polymer, which polymer is chosen from the group comprising a polymer which is naturally present in human tissue , a synthetic polymer, and a non-human natural polymer .

In one embodiment , the microgels comprise at least one polymer selected from the group consisting of hyaluronic acid, collagen, chondroitin sul fate , heparan sul fate , keratan sul fate and dermatan sul fate

In one embodiment , interactions of the compound to study and the polymer in the microgels are studied .

In one embodiment , the microgels comprise at least one polymer, which polymer is naturally present in human subcutaneous tissue and where interactions of a drug and the polymer are studied . As shown in the examples , interactions of drugs with polymers present for instance in human subcutaneous tissue can be studied using the system and method of the invention .

In one embodiment , the fluid comprising the at least one compound to study is recirculated . This is suitable for instance for compounds , which are expensive or di f ficult to obtain . In these situations , with very small amounts of a drug substance available it would be possible to use a recirculation valve , that would then enable the same small volume to be recirculated over the gels during the whole run or a part of the experiment .

In one embodiment , the method further comprises the step of : d) washing the microgels by subjecting said microgels to a washing solution, followed by repeating step b) and c) , wherein the at least one compound to be studied optionally is a different compound to be studied, optionally followed by repeating steps b) , c) and d) a plurality of times.

Thereby the microgels can be reused and several different compounds can be studied in one experiment.

In one embodiment the washing solution is a salt solution. In one embodiment the solution is added by manually switching the solution to be added to the chip. In one embodiment the solution is added by automatically switching the solution to be added to the chip using a commercially available switching device.

Examples

Materials and methods

Chemicals. Polydimethylsiloxane (PDMS) was obtained as the kit Sylgard 184 including Elastomer base and curing agent from GA Lindberg ChemTech AB. Picosurf 5% was purchased from Sphere fluidics and Novec 7500 >99% was purchased from 3M. Sodium hyaluronate (100-300 kDa) was purchased from Contipro A.s. and DC Beads (70-150 pm) from BTG international group. The linker N- ( 2-aminoethyl ) acrylamide hydrochloride (AEA) was purchased from abcr, Spectra/Por™ dialysis membrane 3500 Da by Spectrum™ was purchased from Fischer scientific. SUEX photoresist film was purchased from DJ MicroLaminates. 2-propanol was purchased from Merck and Ethanol 99.7% was from Solveco. Acrylic acid (anhydrous 99%) , N, N, N' , N' -tetramethylethylenediamine (TEMED) (ReagentPlus 99%) , N, N' -methylenebisacrylamide (99%) , ammonium persulfate (powder >98%) , sorbitane monostearate (Span 60) , sodium phosphate monobasic

(ReagentPlus >99%) , sodium phosphate dibasic (ReagentPlus >99%) , amitriptyline hydrochloride (A8404) , 1H, 1H, 2H, 2H- perf luoro-l-octanol (>97%) , protamine sulfate (P3369) , cytochrome C from equine heart (C2506) , EDC (N— (3— dimethylaminopropoyl ) -N' -ethylcarbodiimide hydrochloride, lithium phenyl-2 , 4 , 6-trimethylbenzoylphosphinate (LAP) , Sigmacote® sodium chloride (NaCl) , sodium hydroxide (NaOH) , HOBt ( 1-hydroxybenzotriazole hydrate) and Acetonitrile (anhydrous 99.8%) was all purchased from Sigma aidrich.

The microscope used for all microscopy studies was an Olympus BX51 with an Olympus DP73 camera. The pressure pump used for microfluidic experiments was an OBK M111+ with digital flow sensors 3 and 4, all from Elvesys, Paris. The tubing used was a PTFE #30 AWG thin wall tubing natural from Cole Parmer. The UV lamp was a UVP crosslinker CL-1000 from Analytikj ena . A ONM-1 Manipulator, Micropipette puller PN-31, micropipette grinder EG-400, and micro-forge MF-900, all from Narishige, were used to pull and polish and move micropipettes during the micromanipulator-assisted microscopy experiments. A Harrick Plasma cleaner PDC-32 G was used for PDMS bonding to glass.

Microfluidic chip fabrication. The production of microfluidic chips used in this work are manufactured with standard soft lithography techniques.

Masters for the microfluidic chips were produced by first laminating a pre-treated 4-inch silicon wafer with a 210 pm thick SUEX photoresist film . The lamination was done at 60°C followed by 5 minutes post lamination at 80°C . The laminated wafer was exposed to i-line filtered UV-light in eight cycles of 30 seconds with 45 seconds cooling in between each cycle , and then left in 80°C for another 3 hours followed by cooling overnight . Finally the noncrosslinked laminate was washed away "developed" with a development solution for 60 minutes with renewal of the solution after 45 minutes , the development was stopped by rinsing the wafer with isopropanol .

The microfluidic chips where made of PDMS using the Sylgard 184 kit . Sylgard 184 elastomer base and curing agent were mixed 10 : 1 and cast over the master in a 14 cm petri dish . Vacuum was used to remove bubbles and the PDMS was cured for 1 hour at 70°C . After curing, the PDMS was peeled from the master, cut with a scalpel , and holes of diameter 0 . 75 mm for inlets and outlets were punched . Cut PDMS structures were then covalently bound to a glass slide after both were pretreated with air plasma for 30 seconds . Both pieces were then immediately put in contact followed by 1 hour in oven at 70°C .

The microfluidic chip used for production of HA microgels were finally treated with 8 pl of Sigmacote® to produce hydrophobic channels . The microfluidic chip was left to dry in the oven at 80 ° C for 1 hour . Drawings of the microfluidic chips for interaction studies (MIS ) and for droplet production (MDP ) can be seen in figure 1 .

Microgel fabrication . HA microgels and polyacrylic acid

( PA) microgels where both produced in-house while DC beads where purchased from Boston Scienti fic . First , 2 . 6 g of acrylic acid, 11 g of 2 M NaOH (aq) , 6.5 g of 80 mM linker N, N-methylenebisacrylamide and 109 pl of TEMED were mixed with deionized (DI) water to a total volume of 20 ml. Separately, 0.09 g of Span 60 was dissolved in 30 ml cyclohexane for 2 days with stirring. The Span 60 solution was heated to 45 °C and kept under a nitrogen atmosphere at 400 rpm stirring. Finally, 364 pl of 0.18 M ammonium persulfate was added to 10 ml of the reaction mixture which immediately after was injected into the Span 60 solution. The temperature of the emulsion was increased to 60 °C. After stirring at 1000 rpm for 30 minutes, 40 ml of ethanol was added to break the emulsion. The ethanol/water phase containing microgels was transferred to a beaker and mixed with an excess of methanol. After mixing, the methanol phase was removed and the microgel slurry was dried at 60 °C. After drying, the microgels where resuspended in DI water and rinsed 4 times with DI water. Finally, pH was adjusted to be above 9 by adding small aliquots of NaOH solution to keep the microgels stable.

Sodium hyaluronate was first functionalized by converting carboxylic acid groups into acrylamide groups. To this end, 400 mg sodium hyaluronate was dissolved in 50 ml DI water together with the linker AEA (105 mg) ; the solution was stored in the dark. HOBt (152 mg) was dissolved separately in a 1:1 mixture of DI water and acetonitrile during gentle heating. After cooling, the HOBt solution was added to the reaction mixture and the pH was adjusted to pH 6 with IM HC1. After adding, EDC (287 mg) the reaction mixture was stirred for 24 hours at room temperature. The mixture was dialysed (Spectra/Por™ 3,5 kDa membrane) first against 0.6% (w/v) NaCl (aq) solution at pH 3.5 for 24 hours, then twice against DI water with pH adjusted to 3.5 for 24 hours, and finally against pure DI water. After dialysis the reaction mixture was filtered through a filtration paper and freeze dried. The degree of modification was determined by ²H NMR in D₂O indicating that 45% of disaccharide units were modified.

HA microgels were produced by utilizing a droplet-making chip of T-junction geometry (fig. 1) . Freeze-dried HA-Am was dissolved in a 0.1 % (w/w) solution of LAP in DI water to a concentration of 2% (w/w) ; dissolution was done in room temperature overnight. A mixture of Novec 7500™ with 0.5% (v/v) Picosurf™ was prepared and used as the continuous phase. The two solutions were transferred to falcon tubes and connected to the droplet-making chip using tubing with inner diameter (ID) of 300 pm. Flow rates were set to 5 pl/min and 120 pl/min for the aqueous and oil phase, respectively. Droplets were collected into a glass beaker and cross-linked with UV-light at 365 nm and irradiation energy of 1000 pj/cm² for 10 minutes. The emulsion was filtered through a 70 pm filter 3-10 times with DI water until oil and surfactant were washed away. The microgels were then re-suspended in 5 mM PBS (pH 7.4) (called PBS (5) for the rest of this work) and stored in refrigerator .

Determination of microgel volume. Volume of microgels used during this work were calculated from the diameter of each microgel. Images of the microgels were obtained with a Olympus BX51 microscope with an UMPlanFI 5x lens equipped and a Olympus DP73 digital camera connected. Diameter measurements were done with the imaging software cellSens Dimension version 1.7.1 from Olympus Corporation.

Interaction experiments. The MIS shown in Fig. 2 was used for all interaction experiments. Two tubings of inner diameter 300 pm were connected to the microfluidic chip via a flow sensor . A stock solution with an appropriate concentration of the substance under study in a buf fer was pumped into the chip through one of the tubings ( inlet 1 ) and PBS ( 5 ) or PBS solution 5 mM (pH 5 . 9 ) through the other ( inlet 2 ) . The flow rates were controlled with the Elveflow smart interface software . By varying the flow rates of the two solutions , it was possible to acquire di f ferent concentrations in the chip without having to manually change solutions . This also made it possible to automate longer experimental series . When inlet 1 and 2 had been connected a third tube was attached for loading of the microgels . At the start of each experiment PBS was first degassed for 1 hour and then pumped through the chip before manually loading the traps with microgels using a syringe connected to a separate inlet . When the microgels were in place , the flow rates of inlet 1 and 2 were modi fied to get the desired concentration of drug substance flowing around the trapped microgels . The total flow rate through the chip was 200 pl/min i f nothing else is written . Both single concentration experiments and experiments with concentrations changing over time were performed .

Physicochemical properties of microgels and drug substances .

To understand and interpret the results presented in this work it is necessary to have knowledge about some physicochemical properties of the microgels and drug substances used in the experiments . Table 3 shows charge concentration and separation of charges along the polyelectrolyte chains (b ) in the three microgels used; all data are for microgels in 5 mM salt solutions (pH 7 . 4 ) . The microgels were chosen to cover a range of properties . PA microgels consist of covalently crosslinked sodium polyacrylate . The chains are flexible , highly charged and interact mainly electrostatically with proteins and micelles . HA microgels are made of covalently crosslinked sodium hyaluronate chains . The HA backbone is fairly hydrophilic but bulkier and less flexible than PA. The b value given in Table 3 is calculated taking into account that 45 % of the carboxylate groups are modi fied by the crosslinker . The linear charge density is thus considerably lower than for PA, which explains the low charge concentration in the microgel . DC beads consist of uncharged polyvinyl alcohol ( PVA) chains crosslinked with negatively charged poly ( 2-acrylamido-2- methylpropanesul fonate ) chains ( PAMPS ) . The linear charge density of PAMPS is the same as for PA and both are expected to interact stronger than HA with proteins and micelles . However, the contribution to the osmotic pressure from the PVA chains prevents the network to collapse to the same extent as PA networks upon loading of proteins and micelles . For HA, the large molecular weight per charge , is expected to play a similar role as PVA in DC bead .

In Table 4 charge and si ze properties of the three drug model substances used are presented . Just as for the microgels , the substances were chosen to cover a range of properties . We have chosen cytochrome C ( CytC ) as model of globular proteins . Previous investigations show that its net charge is suf ficient to make it associate with negatively charged polyelectrolytes at moderate ionic strength but not large enough to induce volume phase transition in PA networks . Furthermore , its propensity to sel f-associate is small . Protamine is a chosen as a model of a highly charged peptide drug . The biological function of this linear peptide is to condense DNA in sperm heads, but is known to associate very strongly with negatively charged polyelectrolytes in general. Amitriptyline is used as model of micelle-forming amphiphilic drugs. Previous studies show that the micelles associate with negatively charged polyelectrolytes and induces volume phase transition in PA networks and DC bead. The dimension of the micelles is similar to that of a CytC molecule but the net charge is much higher, making them suitable for comparing charge effects. The micelle size in aqueous solutions was recently determined with small-angle x-ray scattering. The data for AMT in Table 3 are representative for concentrated AMT solutions, comparable to the concentration inside fully loaded microgels.

Table 1: Charge concentration of microgels and charge density of linear charged polymers making up the polymer network of the microgels .

Table 2: Charge density, molecular weight, charge of one molecule and size (length for PRO, radius for CytC and AMT micelles) . Charge density and radius presented for AMT is the values for micelles formed at approx. 40 mM (CMC) in PBS (5) .

1,2: Values for micelles of AMT (aggregation number 40)

Microgels on a chip

Below we will demonstrate that valuable information about the interaction between macromolecular drugs and polyelectrolytes can be obtained from microfluidics studies of polyelectrolyte microgels. Clearly, protamine gave rise to a larger volume decrease than CytC. As another example, AMT leads to a dramatic volume collapse of HA microgels. In subsequent sections we will first show how to characterize the inherent responsiveness of the microgel networks, and after that we will describe different types of interaction studies. Results from an investigation of the mass transfer rate to the microgels in the microfluidic traps, showing that conditions were comparable to those in the micromanipulator setup used earlier and that the confinement to the trap did not affect the swelling properties.

Characterization of microgel networks: Osmotic stress For a given type of polyelectrolyte, both the elasticity of the network and the electrostatic coupling between the drug and the network determine the magnitude of the volume response . To be able extract information about the latter, it is important to characteri ze the former in the absence of drug . This can be done by exposing the microgel to osmotic stress . A convenient method is to monitor the volume response to variation of the ionic strength in the solution . Fig . 3A shows the volume response of the microgels of the present study to increasing concentrations of NaCl in the solution . The ordinate shows the actual microgel volume V divided by the volume V_o in a reference solution containing 5 mM PBS (pH=7 . 4 ) , present as a background in all solutions . Fig . 3B shows theoretically calculated response curves for networks of di f ferent values of the elasticity parameter p and fraction of charged segments in the polyelectrolyte chains q. The former can be interpreted as the apparent number of chain segments between crosslinks , a number decreasing with increasing degree of crosslinking of the network chains . In the model , the osmotic swelling forces are balanced by the contractive elastic forces in the network . The former is largely determined by the di f ference in mobile ion concentration between the microgel and the liquid . According to theory the ef fect produced by adding salt increases with both increasing p and q. The di f ference between the HA and PA networks in Fig . 3A could thus result from both di f ferent degree of crosslinking and linear charge density of the chains . In DC bead, the PAMPS component has the same linear charge density as PA, but a substantial part of the network consists of uncharged PVA chains . It is likely that the additional swelling pressure from the PVA chains contributes to the di f ference between DC bead and PA, but there could also be a di f ference in degree of crosslinking . For all three microgels , the deswelling ef fect per amount of added salt is largest at low salt concentration where only small amounts of salt enters the microgel ( Donnan ef fect ) . In the model this is purely an ef fect of the decreased swelling pressure in the microgel due to the increased osmotic pressure in the solution, allowing the network to relax to a state of lower elastic energy . The ef fect of increasing the salt concentrations was small for concentrations above 100 mM . In this range the swelling pressure due to the non-uni form partitioning of mobile ions , created initially by the presence of fixed charges on the network, has been removed . The amplitude of the volume change caused by the addition of salt is a measure of the responsiveness of the network . As indicated above , one motive behind the drug-induced volume decrease is the elastic response to the reduced swelling pressure deriving from the replacement of network counterions by macromolecular drugs (proteins , micelles ) . Another is the attractive electrostatic forces between the network and the macromolecules . As a means to distinguish between those we will use the V/V_o value at 150 mM NaCl in drug free solutions as a reference level .

Drug - polyelectrolYte interactions

In this part we evaluate a set of microfluidics experiments designed to provide information about the nature and strength of drug - polyelectrolyte interactions . After confining the microgels to traps , we always first pumped PBS solution through the chip to determine the equilibrium volume V_o of the microgels in a drug- free reference state . I f nothing else is stated, the PBS solution had an ionic strength of 5 mM and pH=7 . 4 . We then changed to the drug solution of choice and recorded how the microgels developed with time . Fig . 4 shows examples of how PA, HA, and DC bead microgels responded to PRO and CytC solutions, respectively. Since the solution in contact with the microgels was constantly renewed, each microgel was, in practice, in contact with an infinite bath of drug solution ("reservoir") . After a sufficiently long incubation period the volume of the microgel relaxed to a new level, where the microgel was considered to be in equilibrium with the solution reservoir. The microgel volume in this state will be referred to as V_end -

In the first section below we demonstrate what information can be extracted from the equilibrium properties V_o and Vend - After that we demonstrate what information can be obtained from kinetic properties, i.e., the rate of change from V_o to V_end, and from observations of the internal structure of microgels during that process. Finally we evaluate a "fast track" type of experiment where the setup was programmed to increase the concentration in the solution in steps after regular time intervals.

Equilibrium properties

Protein drug binding . The swelling profiles for PA microgels in PRO solutions (Fig. 4A) and CytC solutions (Fig. 4B) show that PRO induces the largest volume decrease at long times. The result is in agreement with equilibrium data from the literature, and can be attributed to the difference in protein net charge (Table 1) . Fig. 3A shows swelling isotherms calculated from theory for a model network

in equilibrium with solutions of spheres (r_s=lnm) of opposite charge to the network and 5 mM salt. The microgel volume per network charge is plotted vs. the concentration of spheres in the solution for spheres of different charge number. Fig. 3B shows the corresponding binding isotherms, where ft is the loading level expressed as protein/network charge ratio . Clearly, a protein of charge +21 ( PRO) is expected to give rise to a much larger contraction of the network than a protein of charge +7 ( CytC ) at a given protein concentration in the solution . The calculations also show that a protein of charge +21 is expected to have reached the maximum loading level at the lowest concentration in the experiments in Fig . 4A, explaining why Vend did not change when the concentration increased from 10 to 250 pM . For CytC V_end decreased with increasing concentration in the investigated range ( Fig . 4B ) . This is in agreement with results from a previous study by means of the micropipette technique , where a minimum in microgel volume was observed at

« 0 . 8 . The ef fect can be attributed to excluded- volume repulsion between the protein molecules , as was shown by theoretical calculations with a more detailed version of the microgel model . This can also explain the minimum in the swelling curve at 250 pM CytC in Fig . 4B, since it is likely that the final loading level at this concentration exceeded 0 . 8 .

Turning to HA microgels , it is clear from Figs . 4C, D that PRO caused a much large volume contraction than did CytC at 10 pM protein in the solution . Increasing the PRO concentration did not signi ficantly change V_end - For CytC, the volume decreased substantially as the concentration was increased to 50 pM . The result shows that for PRO the maximum loading level was reached well below 10 pM but for CytC the loading level increased progressively in the investigated concentration range .

Figs . 4 E and F show that the di f ference between the two proteins remained also for their interactions with DC bead . The results were qualitatively similar to those for PA. Thus , PRO produced the largest volume change and both proteins reached substantial loading levels already well below 10 pM . Furthermore , V_end for CytC increased somewhat as the concentration increased from 10 to 250 pM, suggesting an excluded volume ef fect similar to that in PA microgels . ( Careful inspection shows that also PRO gave rise to a similar but smaller re-swelling of DC bead and HA but not PA microgels . )

The Vend /V_o ratio determined for the di f ferent systems are presented in Table 5 . Each entry is based on the average for eight microgels . Comparison shows that PRO interacts stronger than CytC with all three networks investigated, and that this can be explained by PRO' s higher charge number ( cf. Fig . 19 ) . One maj or factor behind the ef fect is that the gain in entropy from replacing the network counterions with protein increases dramatically with the charge number of the protein . This has a huge ef fect on the concentration of protein in the solution in equilibrium with a gel and thus largely determines in what concentration range the maj or volume change takes place .

In principle , the ef fect does not require intimate contact between the protein and the polyelectrolyte chains . However, with increasing charge density on both, the attraction between them increases , and will in such cases increase the driving force for protein binding to gels . Apart from contributing to the shi ft of the swelling isotherms to lower concentrations , the attraction will increase the negative slope of the isotherms and can give rise to a volume phase transition (VPT ) from the swollen to a much collapsed state ( cf. Fig . 19 ) . The very low Vend/Vo indicates that PRO induced such a collapse in all three microgels . In contrast , V_end/V₀ for CytC were considerably larger in all cases , and the protein did not induce a VPT . To demonstrate that , we show in Fig . 5 equilibrium swelling isotherms for HA microgels in solutions of PRO and CytC, respectively, recorded with the microfluidic set up . Clearly, the swelling decreased gradually with increasing concentration of CytC in the solution but the microgel never reached the fully collapsed state of the PRO system . This shows that CytC was attracted to HY but did not form dense complex phases as did PRO .

Amphiphili c drug binding . As a complement to the protein studies we demonstrate that the method can be used to study also the interaction between polyelectrolytes and small molecules , exempli fied here by the cationic amphiphilic drug AMT . Fig . 6 shows a swelling isotherm for HA microgels as a function of the AMT concentration in the liquid solution ( 5 mM PBS ) . The microgel volume decreased gradually with increasing AMT concentration until j ust above 10 mM where a discontinuous transition to a much collapsed state took place . The behavior is in qualitative agreement with previous results for AMT interacting with other polyelectrolyte networks . AMT is known to form globular micelles in concentrated aqueous solutions with dimensions similar to the CytC molecule . However, the micelle charge is considerably larger than the net charge of a CytC molecule ( cf. Table 3 ) . This explains why the microgel volume (V/V_o ) above CCC for AMT is comparable to Vend/Vo of PRO rather than CytC . Fig . 7 shows time profiles for the volume response of in 23 mM AMT solution . The microgel volume decreased rapidly down to the equilibrium value , as expected since the concentration was well above the critical collapse concentration ( cf. Fig . 6 ) . Drug and polyel ectrolyte concentra ti on in mi crogel . The results presented so far show that it is straightforward to rank the drug molecules after increasing binding strength to a given microgel based on V_end/V₀ values . For example , for binding to HA microgels we have CytC < PRO < AMT (micelle ) , following the order of increasing charge number . However, equilibrium swelling data can also be used to rank the microgel networks based on their strength of interaction with a given drug . Here , it would be most informative to compare swelling isotherms showing V/V_o as function of the protein concentration in the solution ( cf. Fig . 8 ) . However, for highly charged proteins , such as PRO, the binding may start at so low concentrations ( cf. Fig . 19 ) that it is practically di f ficult to establish isotherms that cover the full dynamic range . Here we focus instead on the information that can be obtained from comparison of V_end/V₀ values . Since V_o is determined by several factors , including linear charge density, degree of crosslinking and the concentration of polymer in the state where the network was created, it is essential to convert V_end/V₀ into a quantity showing the density of the end state . Here we used the concentration of network charges in the reference state C_o ( Table 3 ) to calculate the concentration of network charges Cp ^ld = C₀V₀/V_end) and tration of protein molecules and micelles C™^d ( =

the end state of the microgels . In all cases we assumed ? = 1 . Previous investigations show that deviations from that value should not be larger than 10 % at the present ionic strength . The result is given in Table 5 . For comparison we have included data for the microgels in 0 . 15 M NaCl solutions without drug present . The first thing to notice, is that the Cp^nd values followed the order PA>DC bead> HA for all three model drugs. It highlights the importance of correcting for the differences in C_o, since DC bead showed the highest V_end/V₀ but not the lowest Cp^nd . Second, we note that PRO and AMT transformed the PA microgels into a very dense state. For example, the C™^d value for AMT micelles corresponds to a micelle volume fraction of around 0.5, meaning that the network charge concentration in the space between the micelles was ca. 3.6 M. This shows that there are strong interactions between the components. Theoretical investigations have revealed that, for packing densities of that magnitude, the swelling pressure from the excluded volume repulsion, balanced by the attractive electrostatic force, largely determines the swelling equilibrium. As already noted, CytC induced a much smaller contraction of the PA network. The C™^d value corresponds to a volume fraction of ca. 0.3 which is still rather high. However, in this case it is the elastic network forces that brings the protein molecules together (ref) . This explains why the contraction of the network induced by CytC loading in 5 mM PBS was comparable to the effect of adding 0.15 M NaCl (Table 5) .

The polyelectrolyte chains in PA and DC bead microgels have the same linear charge density and we could have expected them to behave more similar. However, the comparatively large swelling of DC bead can be attributed to the neutral PVA chains providing an extra swelling pressure preventing the microgels to collapse to the same extent as the PA microgels. For PRO, AMT, and CytC, Cp^nd was respectively 4.6, 4.7 and 2.4 times lower for DC bead than for PA. The high values for PRO and AMT are explained by the crowded state produced by these molecules, where introduction of PVA chains in the aqueous regions between the proteins/micelles is expected to have a relatively large effect on the swelling. For CytC, there is more space available, and so the effect of incorporating PVA is larger. The same argument explains the difference between PA and HA in this respect. The lower linear charge density of HA means a higher molecular weight per charge. Thus, at a given network charge concentration the polymer concentration in the microgel is larger for HA than for PA.

PRO induced the strongest contraction of PA and DC bead microgels. This can be attributed to a combination of strong electrostatic attractions to the networks chains and a small diameter of the peptide chain, which in its extended conformation can be efficiently packed. HA microgels were more contracted by AMT than by PRO. This reversal of order may be attributed to the large difference in linear charge density between PRO and HA ("charge mismatch") .

Table 3: Minimum volume ratio (Vmin/VO) of PA microgels, HA microgels and DC beads at exposure to a high concentration of CytC, PRO or AMT. Volume presented are the average of 8 microgels if nothing else i stated. C_nc, (max) are the network charge concentration in the microgel at V_min, C_mc, _(max) are the concentration of drug substance or micelle in the microgel at V_min(p=l) .. C_nc, _(max) CoVo/V_min where Co is the network charge of the microgel in pH 7.5 mM 5 M NaCl before experiment. C_mc, _(max) = C_nc, _(max)/Z where Z is the charge of one drug substance/micelle (7 CytC, 21 PRO, 40 AMT micelle) .

Effect of salt.

Addition of salt destabilizes complexes between proteins and polyelectrolytes and at sufficiently high ionic strength the protein is expected to be released. In the present method the release is demonstrated directly as decolorization of the microgel if the protein is colored or fluorescent, or indirectly if the release is accompanied by microgel volume change. In the latter case, differences in the interaction strength can give rise to qualitative differences in the volume response. This is illustrated in Fig. 9showing how the volume of HA, PA and DC bead microgels in 250 pM CytC solutions (5 mM PBS, pH 7.4) varies with the concentration of NaCl added to the solution. With no NaCl added, the microgels where compact and contained substantial amounts of protein. With increasing NaCl concentration, the HA microgels increased in volume up to a certain point, thereafter the volume decreased . The behavior is explained by swelling of the network during the phase where maj ority of the protein molecules were released, followed by deswelling due to the common salt ef fect on the swelling of charged networks ( cf. Fig . 3A-D) . For PA and DC bead the swelling decreased monotonically in the entire salt concentration range . The di f ference compared with HA can be explained by CytC binding stronger to the PA and DC bead networks , leading to a slower release of the protein as function of salt concentration . The relationship between swelling ( solid curves ) and binding ( dotted curves ) is illustrated by the theoretical model calculations in Fig . 10 , comparing the behaviors of two network with strong and weak attraction to the charged spheres , respectively . Weak attraction was modelled implicitly by adding an excess free energy per bound sphere . In both cases large amounts of spheres bind at low salt concentrations , but in the weakly attracting system the amount decreases rapidly with increasing salt concentration and is practically zero for salt concentrations larger than 20 mM . Under the latter conditions the release is initially accompanied by swelling due to the replacement of spheres by monovalent counterions . However, when most of the spheres are released the swelling decreases again due the increased osmotic pressure from the bulk solution . When the attraction is strong, the released amount is never large enough for the swelling tendency to overcome the deswelling induced by the common salt ef fect , and so the swelling decreases slowly but monotonically .

Conclusions . First of all it is clear that all systems of microgel-drug substances in this work reach equilibrium within a timescale that is long enough to be able to follow with optical microscopy methods , but at the same time not longer than what would feasible to measure (hours instead of days ) . The timescale can be shi fted to a certain degree for a given system though i f needed, by changing the si ze of the microgel , the concentration of drug substance or flow rate . How big of a shi ft is heavily depending on type of microgel and drug substance , and need to be tested for each system, but from the results presented in this work it can vary from seconds up towards 20 minutes .

Amphiphilic drug loading and release

Loading . Fig . 11 shows time profiles for the volume response of DC bead microgels in 0 . 52 mM AMT solutions in 5 mM PBS . Qualitatively the behavior resembles that of DC bead + PRO ( cf . Fig . 4E ) , including the fact that the microgels displayed a core-shell phase coexistence during the loading process . However the results show that , after a short lag period where the volume changes very little , the initial deswelling rate of DC beads is very quick compared to when exposed to PRO .

Rel ease . Fig . 12 shows how the volume of HA microgels preloaded with AMT changed as a function of time in 5 mM PBS containing 0 . 15 M NaCl . Since the release medium contained no AMT the liquid flowing through the chip acted as a sink . After a lag period, the microgel abruptly increased in volume , indicating that AMT was released . The volume V_o determined for the same microgel in equilibrium with the release medium was not fully recovered . In principle , it could indicate that not all AMT was released or that the swelling of the network was not completely reversible ( see below) . For systems where uncertainties regarding the latter can be excluded, experiments of this type can thus be used to investigate the reversibility of the interaction between drug molecules and polyelectrolytes . The swelling rate can provide information about the how the interaction with the polyelectrolyte af fects the di f fusivity of the drug molecules , and how the transport properties are af fected by variation of the ionic strength .

Release properties are , of course , of special importance in the study of microgels intended for drug delivery . The swelling rate was directly coupled to the release which was rate controlled by the di f fusion of drug monomers through the depletion layer . More precisely, the release/ swelling rate was directly related to the concentration gradient in the depletion layer, determined by the thickness of the depletion layer and the local concentration of free drug monomers in equilibrium with the micelles at the core boundary . The latter quantity is a measure of the stability of the drug - polyelectrolyte complexes in the core and therefore a measure of the strength of interaction between the drug and the polyelectrolyte . In principle , this makes it possible to rank di f ferent drug molecules in order of interaction strength by comparing their swelling rates . However, it would be necessary to normali ze the result with their individual di f fusion coef ficients in water .

A phenomena that can be seen after release of PRO and AMT from HA microgels ( Fig . 13 , 29 ) is that the microgels do not return to the volume they had before binding of the drug substances . In some cases there are also a clear deformation of the spherical shape of the microgels ( Fig . 14 ) . This indicated structural damage of HA microgels when exposed to high concentrations of drug substances with very high charge densities , and is something that will be important to consider .

In all systems (microgel-drug substance ) the initial si ze of the microgel will also have an ef fect on the kinetic of deswelling . The larger the microgel are the higher the number of charges in the microgel is which means that a larger amount of drug molecules need to di f fuse into the microgel and to reach p=l . This in turn means that it takes longer time to reach p=l and V_min/V₀.

The experimental procedure of obtaining the presented data went smoothly with a mostly automated processes . One factor to consider though is that during the collapse/deswelling of microgels the spherical shape of the PA and HA microgels exposed to PRO and AMT are deformed ( Fig . 15 ) , the shape is recovered when reaching a plateau in volume change or V_min/V0 . But the volume measurements during the collapse is in these two cases estimations and not exact volumes . The same deformation was not seen when exposed to the lower charged CytC, and the DC beads always retained its spherical shape .

Release experiments

Release of CytC and PRO from the three microgels at di f ferent NaCl concentrations ( DC beads , HA microgels and PA microgels ) and release of AMT from HA microgels at 0 mM NaCl were investigated . When multiple NaCl concentrations were used during one experiment the concentration was changed every 10 minute going from 0 or 10 mM to 150 or 1 M .

V0 in these presented experimental data is the volume of the microgel before exposure to drug substances in the highest NaCl concentration ( 150 or 1000 mM) . Full release needs to be confirmed with a visual inspection of the microgels since the volume do not give the full picture .

Before discussing the swelling properties of the microgels it is important to state that independent of system ( drug substance-microgel ) complete release of the drug substance will always occur as long as the solution around the microgel can be considered a sink with any amount of ions however small . The rate is heavily af fected by the ionic strength of the solution though, since the salt ions will rapidly equilibrate between the microgels and the solution, and the concentration of salt will be almost the same outside and inside the microgel . The higher the concentration of salt ions are in the microgel the more the electrostatic coupling between drug substance-microgel is screened and the higher the mass transport rate of the drug substances out of the microgel will be . The system properties of the drug substance-microgel system then determine how much a certain ion strength af fect the release of drug substances and swelling of the microgels . Where higher charge density in the drug substances and higher charge concentration in the microgel network means that a larger amount of ions is needed to screen the electrostatic attraction forces . This screening ef fect are seen in the model in Fig . 9 , where p lowers at increasing ionic strength and a volume change is seen, in this case an increase followed by a decrease in volume , but the volume change is dependent on the system .

The impact of charges on microgels/drug substances on release is clear when comparing the release from PRO loaded and CytC loaded microgels . When PRO loaded PA microgels and DC beads both with high charge concentration ( 90 and 60 mM) are exposed to increasing NaCl concentration the volume change is minimal , with no swelling at all until 1 M when both microgels swell a few percentage (swelling of PA microgels at 1 M not shown) , but no visual release of PRO is observed (Fig. 16) . When PRO loaded HA microgels with a lower charge concentration (10-20 mM) is exposed to increasing NaCl concentrations, swelling and visual release of PRO 0 starts at 10 mM with full release at 150 mM NaCl (Fig. 13) .

CytC loaded PA microgels and DC beads start visually releasing CytC at 10 mM NaCl and close to full release is seen at 150 mM (Fig. 17,18) . The microgels do not swell, DC beads shrink further at increased NaCl concentration, while PA microgels stay close to the same volume ratio of 1 through the experiment. This is expected. The volume change at release follows the results seen in table 5, where V_min/V₀ at exposure to CytC (250 pM) is the same as the volume at 150 mM NaCl for PA microgels and slightly higher for DC beads. CytC loaded HA microgels start visually releasing most of the CytC at exposure to only PBS (5) and the volume increases several times before starting to decrease again, the release are estimated to be complete at 50 mM NaCl exposure by visual examination (Fig. 17) .

Interactions experiments - summarization

It has been shown that the MIS in combination with volume measurement through microscopy can be used for interaction studies with both different types of microgels and drug substances. Ranging from the highly charged peptide PRO (4.25 kDa, +21 net charge at pH 7.4) to a large enzyme with low charge density CytC (12.3 kDa, +7 net charge at pH 7.4) , to a small amphiphilic molecule AMT forming micelles (277 g/mol, +1 net charge) . It was possible to measure both binding and release of these drug substances to the three types of microgels used ( DC beads , PA and HA microgels ) with some limitations in volumes measurements of the microgels during collapse , where the volume in some cases being estimations .

When investigating release there are also a limitation in possible structural damage of HA microgels as the results of very strong interactions with AMT and PRO . This may make it di f ficult to compare the volume after release to the starting volume before binding and release of a PRO or AMT . Which in turn makes it di f ficult to determine with 100% accuracy when full release are achieved, but combining visual examination of microgels with the volume should give a good approximation . It is important to remember that the deformation and structural damage was only seen for the very highly charged molecules with a charge density far above most peptides and proteins , and not for CytC for example .

Claims

Claims

1. A microfluidic system comprising a. a microfluidic device and, b. a plurality of polymer microgels, wherein the microfluidic device comprises:

- at least one fluid inlet (1,2) ,

- at least one fluid outlet (6) ,

- at least one main channel (5) , in fluid contact with the fluid inlet (1,2) and the fluid outlet (6) ,

- a plurality of hydrodynamic traps in fluid contact with the at least one main channel (5) , wherein each hydrodynamic trap is an open cavity with a size and a shape which are adapted to trap and accommodate one microgel, wherein the resistance for a microgel flowing in a liquid in the main channel (5) to enter an empty hydrodynamic trap from the main channel (5) is lower than the resistance for the microgel to continue in the main channel (5) , and wherein the resistance for a microgel to enter an occupied hydrodynamic trap is higher than the resistance for the microgel to continue in the main channel (5) , wherein each hydrodynamic trap has an inlet and at least one outlet, wherein the outlet is downstream of the inlet in the at least one main channel (5) in an intended direction of flow in the main channel (5) .

2. The microfluidic system according to claim 1, wherein the microgels comprise at least one polymer, which polymer is naturally present in human subcutaneous tissue . 54 The microfluidic system according to claim 1 , wherein the microgels comprise at least one polymer, which polymer is chosen from the group comprising a polymer which is naturally present in human tissue , a synthetic polymer, and a non-human natural polymer . The microfluidic system according to any one of claims 1-3 , wherein the microgels comprise at least one polymer selected from the group consisting of hyaluronic acid, collagen, chondroitin sul fate , heparan sul fate , keratan sul fate and dermatan sul fate . The microfluidic system according to any one of claims 1-4 , comprising a programmable controller communicatively coupled to a camera with an optical system capable of imaging the si ze of the microgels , the controller being adapted to receive an image of the microgels and calculate their volume based on the image , the controller being communicatively coupled to a pump and able to control the flow rate and optionally the concentration of a compound to study, the controller being adapted to calculate and present a result related to interactions between a microgel and the compound . The microfluidic system according to any one of claims 1-5 , wherein each trap has at least one outflow channel , fluidly connecting the trap to a part of the main channel located downstream of the trap, wherein the width of the at least one outflow channel is smaller than the width of the main channel . The microfluidic system according to claim 6, wherein each hydrodynamic trap has at least two outflow channels . The microfluidic system according to any one of claims 1-8, wherein the cross section of the traps have a semi-circular shape. The microfluidic system according to any one of claims 1-9, wherein the main channel (5) comprises a plurality of turns, providing a plurality of channel sections, wherein each section is fluidly connected to a downstream section through at least one trap and the at least one outflow channel of that trap. . The microfluidic system according to any one of claims 1-10, comprising a mixer (3) said mixer (3) being positioned in fluid contact with at least one main channel (5) , the mixer (3) optionally comprising at least one filter, wherein the mixer comprises at least one flow path wherein the flow path comprises a plurality of features selected from obstacles and turns in the fluid flow path. . The microfluidic system according to claims 11, wherein the plurality of features is selected from the group consisting of diverting and re-joining channels, flow splitters, sharp edges of the channel, re-occurring width change of the channel, posts and bifurcations . . The microfluidic system according to any one of claims 1-11, wherein the system comprises a separate 56 inlet (4) adapted to add microgels, wherein the separate inlet is before the plurality of hydrodynamic traps and after an optional mixer (3) , wherein before and after are in relation to an intended direction of a flow from the at least one fluid inlet (1,2) to the at least one fluid outlet (6) .

The microfluidic system according to any one of claims 1-12, wherein the plurality of hydrodynamic traps is at least 20 hydrodynamic traps. . A method for studying interactions between polymers in microgels and other compounds, the method comprising the steps of: a. providing a microfluidic system according to any one of claims 1-12, b. adding at least one fluid to the at least one fluid inlet (1,2) , wherein the fluid comprises at least one compound to study, c. measuring changes in the diameter of the microgels upon or after exposure to the compound to study. . The method according to claim 14, wherein the microfluidic system in step a) is provided by loading a microfluidic device with microgels of a size adapted to the size of the hydrodynamic traps, by adding the microgels suspended in a fluid to an inlet for microgels (4) adapted to add microgels. The method according to any one of claims 14-5 , wherein the microgels are loaded during a constant flow of liquid . The method according to any one of claims 14- 16 , wherein two liquids are mixed, one being a solvent or a buf fer, and the other being a stock solution comprising the compound to be studied to a desired concentration in step b ) . The method according to any one of claims 14- 17 , wherein the concentration of the compound to be studied is changed in a gradient . The method according to any one of claims 14- 17 , wherein the composition of the at least one fluid is changed in a gradient . The method according to any one of claims 14- 19 , wherein the concentration of the compound to be studied or the composition of the at least one fluid is changed by means of a controller . The method according to any one of claims 14-20 , wherein measured data are stored and processed in a programmable computer system . The method according to any one of claims 14-21 , wherein the microgels comprise at least one polymer, which polymer is naturally present in human subcutaneous tissue and where the system is utili zed as a model system for subcutaneous tissue . The method according to any one of claims 14-22, wherein the microgels comprise at least one polymer, which polymer is chosen from the group comprising a polymer which is naturally present in human tissue, a synthetic polymer, and a non-human natural polymer. The method according to any one of claims 14-23, wherein the microgels comprise at least one polymer selected from the group consisting of hyaluronic acid, collagen, chondroitin sulfate, heparan sulfate, keratan sulfate and dermatan sulfate. The method according to any one of claims 14-24, wherein interactions of the compound to study and the polymer in the microgels are studied. The method according to any one of claims 14-25, wherein the fluid comprising the at least one compound to study is recirculated. The method according to any one of claims 14-26, further comprising the step of: d) washing the microgels by subjecting said microgels to a washing solution, followed by repeating step b) and c) , wherein the at least one compound to be studied optionally is a different compound to be studied, optionally followed by repeating steps b) , c) and d) a plurality of times.