Rosette-induced separation of T cells by acoustophoresis

I. INTRODUCTION

Section:

ChooseTop of pageABSTRACTI. INTRODUCTION <<II. MATERIALS AND METHODSIII. RESULTSIV. DISCUSSIONV. CONCLUSIONSSUPPLEMENTARY MATERIALREFERENCESPrevious sectionNext sectionRecent developments in autologous cell therapy have achieved dramatic success in treating blood cancers, and similar therapies are expected to follow for other diseases.11. M. Sadelain, I. Rivière, and S. Riddell, “Therapeutic T cell engineering,” Nature 545(7655), 423–431 (2017). https://doi.org/10.1038/nature22395,22. D. Mai, N. C. Sheppard, and B. L. Levine, “Advances in engineering and synthetic biology toward improved therapeutic immune cells,” Curr. Opin. Biomed. Eng. 20, 100342 (2021). https://doi.org/10.1016/j.cobme.2021.100342 This success has brought to light the urgent need for efficient and affordable processing methods as patient cells are manipulated through many steps in the manufacturing facility. For example, in current chimeric antigen receptor T cell (CAR-T) therapies, manufacturing begins with the isolation of T cells from leukapheresis product collected from the patient, and these cells then undergo downstream modification by viral transduction and expansion before they are re-administered to the patient.33. X. Wang and I. Riviere, “Clinical manufacturing of CAR T cells: Foundation of a promising therapy,” Mol. Ther. Oncolytics 3, 16015 (2016). https://doi.org/10.1038/mto.2016.15,44. O. L. Reddy, D. F. Stroncek, and S. R. Panch, “Improving CAR T cell therapy by optimizing critical quality attributes,” Semin. Hematol. 57(2), 33–38 (2020). https://doi.org/10.1053/j.seminhematol.2020.07.005 In this first purification step, the established workflow involves magnetic bead-based separation, where paramagnetic micro/nanoparticles functionalized with desired antibodies are incubated with the blood product before it flows through a magnetic trap to separate bead-bound from unbound cells. While this method provides an acceptable solution, it requires a supply of clinical grade magnetic beads and it suffers from variable and unsatisfactory yield.55. K. Aleksandrova et al., “Functionality and cell senescence of CD4/ CD8-selected CD20 CAR T cells manufactured using the automated CliniMACS Prodigy® platform,” Transfus. Med. Hemother. 46(1), 47–54 (2019). https://doi.org/10.1159/000495772,66. D. Lock et al., “Automated manufacturing of potent CD20-directed chimeric antigen receptor T cells for clinical use,” Hum. Gene Ther. 28(10), 914–925 (2017). https://doi.org/10.1089/hum.2017.111As an alternative to magnetic separation, microfluidic acoustic separation separates cells based on their physical properties like size, density, and compressibility. In the past two decades, acoustophoresis technology has undergone intense development for biomedical applications in research use, ranging from sample preparation for analysis, to tissue engineering and cell culture, to cell and particle purification.77. J. Nilsson et al., “Review of cell and particle trapping in microfluidic systems,” Anal. Chim. Acta 649(2), 141–157 (2009). https://doi.org/10.1016/j.aca.2009.07.017,88. Y. Gao et al., “Acoustic microfluidic separation techniques and bioapplications: A review,” Micromachines 11(10), 921 (2020). https://doi.org/10.3390/mi11100921 Specifically for separation of blood components without antibodies (i.e., “label-free”), we and others have demonstrated fractionation of leukocyte subsets from mononuclear cells, enrichment of lymphocytes from various blood products, platelet depletion, enrichment of hematopoietic stem cells, and so forth.9–159. R. Dubay et al., “Scalable high-throughput acoustophoresis in arrayed plastic microchannels,” Biomicrofluidics 13(3), 034105 (2019). https://doi.org/10.1063/1.509619010. C. Lissandrello et al., “Purification of lymphocytes by acoustic separation in plastic microchannels,” SLAS Technol. 23(4), 352–363 (2018). https://doi.org/10.1177/247263031774994411. A. Mueller et al., “Continuous acoustic separation in a thermoplastic microchannel,” J. Micromech. Microeng. 23(12), 125006 (2013). https://doi.org/10.1088/0960-1317/23/12/12500612. A. Urbansky et al., “Rapid and effective enrichment of mononuclear cells from blood using acoustophoresis,” Sci. Rep. 7(1), 17161 (2017). https://doi.org/10.1038/s41598-017-17200-913. C. Lissandrello et al., “Acoustic microfluidic separation of sickle cell disease erythrocytes,” In AABB Annual Meeting, San Antonio (Wiley, 2019).14. P. Augustsson et al., “Iso-acoustic focusing of cells for size-insensitive acousto-mechanical phenotyping,” Nat. Commun. 7, 11556 (2016). https://doi.org/10.1038/ncomms1155615. Y. Chen et al., “High-throughput acoustic separation of platelets from whole blood,” Lab Chip 16(18), 3466–3472 (2016). https://doi.org/10.1039/C6LC00682E As a result, the acoustic response of blood cells is well characterized1616. K. W. Cushing et al., “Ultrasound characterization of microbead and cell suspensions by speed of sound measurements of neutrally buoyant samples,” Anal. Chem. 89, 8917 (2017). https://doi.org/10.1021/acs.analchem.7b01388,1717. J. M. Dabbi et al., “Label-free measurement of T-cell activation by microfluidic acoustophoresis,” In 2021 21st International Conference on Solid-State Sensors, Actuators and Microsystems (Transducers) (IEEE, 2021). and numerous studies have confirmed that the low acoustic energy required is not harmful to cells.18–2018. W. J. Savage, J. R. Burns, and J. Fiering, “Safety of acoustic separation in plastic devices for extracorporeal blood processing,” Transfusion 57(7), 1818–1826 (2017). https://doi.org/10.1111/trf.1415819. J. Hultström et al., “Proliferation and viability of adherent cells manipulated by standing-wave ultrasound in a microfluidic chip,” Ultrasound Med. Biol. 33(1), 145–151 (2007). https://doi.org/10.1016/j.ultrasmedbio.2006.07.02420. M. A. Burguillos et al., “Microchannel acoustophoresis does not impact survival or function of microglia, leukocytes or tumor cells,” PLoS ONE 8(5), e64233 (2013). https://doi.org/10.1371/journal.pone.0064233 Furthermore, parallel scaling of the microchannels has achieved throughput exceeding 1 ml/min such that the processing time can be comparable to bulk magnetic or centrifugal separations.99. R. Dubay et al., “Scalable high-throughput acoustophoresis in arrayed plastic microchannels,” Biomicrofluidics 13(3), 034105 (2019). https://doi.org/10.1063/1.5096190However, despite these advances, label-free acoustic separation is insufficient to achieve the highly specific isolation needed, for example, in the CAR-T therapy where a T cell output purity of 90% or higher can be required prior to gene modification. This is because the acoustic response of the major lymphocyte subtypes (T, B, NK) is very similar since they have similar physical properties and morphology. Therefore, to enable the use of acoustophoresis in cell therapy isolations, further innovations are needed. Researchers have previously explored antibody conjugated beads to augment acoustic separation,2121. A. Lenshof et al., “Efficient purification of CD4 + lymphocytes from peripheral blood progenitor cell products using affinity bead acoustophoresis,” Cytometry Part A 85(11), 933–941 (2014). https://doi.org/10.1002/cyto.a.22507 but that approach suffers from the drawbacks already inherent in established magnetic separation.In this work, we show for the first time that by cell–cell binding, acoustic separation can achieve high purity and high separation efficiency of T cells, or alternately B cells, from leukapheresis product. Specifically, we exploited rosette formation, achieved with antibodies but without the addition of foreign particles. Rosettes were first described when sheep erythrocytes were mixed with human peripheral blood mononuclear cells (PBMCs) in vitro and erythrocytes spontaneously bound to specific glycoproteins on the surface of lymphocyte subsets, causing “rosette” like structures.2222. D. H. Boldt and J. P. Armstrong, “Rosette formation between human lymphocytes and sheep erythrocytes.: Inhibition of rosette formation by specific glycopeptides,” J. Clin. Invest. 57(4), 1068–1078 (1976). https://doi.org/10.1172/JCI108349 This led to the discovery of specific glycoproteins on lymphocytes, allowing for targeted depletion by tagging cell-specific surface markers.2323. T. Hunig, “The cell surface molecule recognized by the erythrocyte receptor of T lymphocytes.: Identification and partial characterization using a monoclonal antibody,” J. Exp. Med. 162(3), 890–901 (1985). https://doi.org/10.1084/jem.162.3.890 After the development of tetrameric antibody complexes, it became practical to bind two different cell types as desired by conjugating two antibodies to surface markers of the respective cell types.2424. P. M. Lansdorp et al., “Cyclic tetramolecular complexes of monoclonal antibodies: A new type of cross-linking reagent,” Eur. J. Immunol. 16(6), 679–683 (1986). https://doi.org/10.1002/eji.1830160615 Rosette formation was then commercially leveraged to develop proprietary tetrameric antibody cocktails that bind multiple types of unwanted PBMCs to red blood cells (RBCs). In conventional use, this negative selection rosette reagent is applied to whole blood and creates large complexes of PBMCs cross-linked with RBCs, and these complexes are separated by centrifugation with density medium from the lower density unlabeled target cells.

Here, using continuous flow acoustophoresis instead of centrifugation, we successfully applied a commercial rosette reagent kit designed for the negative selection of T cells. Additionally, because we anticipated that acoustic separation should be more selective than centrifugation in separating single cells from cell complexes having only a small number of cells, we applied the reagent to leukapheresis product (aka leukopak) rather than whole blood. This is significant because in a typical leukopak, the number of RBCs available to bind to the unwanted PBMCs is proportionally two orders of magnitude lower than in whole blood.

Our microfluidic acoustic cell separator operates as illustrated in Fig. 1. A polystyrene rectangular microchannel is coupled with an ultrasonic oscillator and mechanically resonated to establish a standing wave in pressure transverse to the flow.1010. C. Lissandrello et al., “Purification of lymphocytes by acoustic separation in plastic microchannels,” SLAS Technol. 23(4), 352–363 (2018). https://doi.org/10.1177/2472630317749944,1111. A. Mueller et al., “Continuous acoustic separation in a thermoplastic microchannel,” J. Micromech. Microeng. 23(12), 125006 (2013). https://doi.org/10.1088/0960-1317/23/12/125006 The cell suspension is pumped through symmetric side inlets, while the buffer solution is pumped through the center inlet in laminar flow. As the cells pass through the acoustic field, they are driven toward the center axial stream. The larger, denser, and stiffer cells migrate the fastest, leaving the less responsive cells in the side streams. In the specific case of rosette treatment of leukopak, we would expect, as observed, that unbound lymphocytes would be collected in the side outlets, while the complexed cell aggregates would be rapidly driven to the center outlet because of their large size and higher density.

We characterized the purity and separation efficiency of T cells in the output product, optimized the acoustic energy in the operating conditions, and investigated the dependence of the output purity on the number of available RBCs. Additionally, as further validation, we tested the same approach for B-cell purification using the appropriate rosette reagent kit. These results suggest that acoustophoresis could provide an alternative to magnetic separation for cell therapy processing, with the advantages of continuous flow and without the need for clinical grade separation particles, instead conjugating cells to other endogenous cells. Moreover, we envision antibody cocktails tailored more efficiently for acoustic separation rather than for centrifugation, which could further reduce the number of RBCs required in T cell isolation and could expand the approach beyond blood to many other cell separation applications.

II. MATERIALS AND METHODS

Section:

ChooseTop of pageABSTRACTI. INTRODUCTIONII. MATERIALS AND METHODS <<III. RESULTSIV. DISCUSSIONV. CONCLUSIONSSUPPLEMENTARY MATERIALREFERENCESPrevious sectionNext section

A. Whole blood and leukopak processing

Fresh leukapheresis product (leukopak) from healthy human donors was purchased from a vendor operating under approved protocols (Hemacare). White blood cells (WBCs), RBC counts, and hematocrit (HCT) were obtained from a hematology analyzer (Sysmex). The required processing volume of leukopak was washed five times in volume of TexMACS media (Miltenyi Biotec) by centrifuging for 10 min at 200g low brake. Washed leukopak was resuspended in TexMACS media at 50 × 106 WBC/ml (manufacturer recommended concentration) and incubated with RosetteSep Immunodensity human T cell enrichment cocktail or B-cell enrichment cocktail (STEMCELL Technologies) for 20 min. As a control, untreated leukopak was processed directly on the acoustic device after washing.

To test the impact of the RBC:WBC ratio, indicated samples were spiked with additional RBCs, obtained from peripheral whole blood of the same donor. The blood was centrifuged at 800g for 10 min, brake off. Plasma and buffy coat were aspirated, and the pelleted RBCs were spiked into washed leukopak prior to the addition of RosetteSep cocktail.

B. Acoustic separation

The microfluidic channel and associated apparatus have been described previously.10–2510. C. Lissandrello et al., “Purification of lymphocytes by acoustic separation in plastic microchannels,” SLAS Technol. 23(4), 352–363 (2018). https://doi.org/10.1177/247263031774994411. A. Mueller et al., “Continuous acoustic separation in a thermoplastic microchannel,” J. Micromech. Microeng. 23(12), 125006 (2013). https://doi.org/10.1088/0960-1317/23/12/12500625. R. Silva et al., “Rapid prototyping and parametric optimization of plastic acoustofluidic devices for blood–bacteria separation,” Biomed. Microdevices 19(3), 70 (2017). https://doi.org/10.1007/s10544-017-0210-3 Briefly, microchannels were fabricated from laminated sheets of general-purpose polystyrene (Calsak Plastics), which was chosen for its prevalence in medical devices, optical clarity, and advantageous acoustic properties. The microfluidic channel (550 μm × 250 μm × 30 mm) and fluidic ports were precision-milled into the top layer of the polystyrene, which was then thermofusion bonded to a cover layer for a total thickness of 2 mm. After connecting medical grade tubing at the inlets and outlets, the channel was affixed with cyanoacrylate adhesive (Loctite) to a lead zirconate titanate (PZT) bulk transducer element, which generates the acoustic standing wave. The transducer was driven by the conventional function generator and amplifier and was monitored with an oscilloscope. A custom thermally controlled aluminum stage was used as a platform to test these devices at a constant 26 °C using an off-the-shelf temperature controller (Arroyo) and thermoelectric element (Laird). Figure 2 shows the mounting configuration.Each device was calibrated in advance for its acoustophoretic performance by measuring the fraction of washed red blood cells (Interstate Blood Bank, Inc.) that were “focused” (i.e., displaced) toward the acoustic pressure node in the center of the channel at a given acoustic energy density, approximated by the voltage supplied to the transducer. During this calibration step, the optimal frequency of actuation (∼575–625 kHz) was tuned visually under bright field microscopy for each device. As previously reported, in the polymer device, the optimal actuation frequency corresponds to much less than a half wavelength across the channel width, assuming the speed of sound in the saline solution; this is in contrast to the behavior of acoustophoresis in hard wall microchannels.11–2611. A. Mueller et al., “Continuous acoustic separation in a thermoplastic microchannel,” J. Micromech. Microeng. 23(12), 125006 (2013). https://doi.org/10.1088/0960-1317/23/12/12500625. R. Silva et al., “Rapid prototyping and parametric optimization of plastic acoustofluidic devices for blood–bacteria separation,” Biomed. Microdevices 19(3), 70 (2017). https://doi.org/10.1007/s10544-017-0210-326. R. P. Moiseyenko and H. Bruus, “Whole-system ultrasound resonances as the basis for acoustophoresis in all-polymer microfluidic devices,” Phys. Rev. Appl. 11(1), 014014 (2019). https://doi.org/10.1103/PhysRevApplied.11.014014 Because of variability among the devices used in these tests, the optimized transducer voltage and frequency varied between runs; however, in each instance, we ensured that the devices performed comparably according to this calibration and that each could achieve at least 90% depletion of RBCs from the side outlet at a sustainable voltage. To ensure minimal deviation of device performance prior to the leukopak separations, calibrations and rosette experiments were completed consecutively for each test.

For the leukopak separations, the sample was loaded into a syringe, gently degassed by retracting the plunger manually with the outlet closed, and then flowed into the device at 100 μl/min into the side inlets. Simultaneously, degassed media were flowed into the device at 400 μl/min into the center inlet. Low-adhesion Eppendorf tubes were used to collect samples. Syringe pumps (Harvard Apparatus) were used to generate this flow, which allowed for the employment of a syringe stir bar (V&P Scientific) to agitate the samples and prevent settling. Transducer frequency was held constant according to the calibration described above. To account for donor variation, different sample treatments, and variable device efficiency, several voltages were tested for each treatment condition. At very low voltages, purification performance is poor as most waste cells are not displaced and remain in the side streams. At excessively high voltages, the recovery and separation efficiency of target cells falls off as both waste and target cells are depleted from the sides. By taking a broad range of voltages, we ensured that a near optimal voltage between these two extremes was found for each sample condition. Practical constraints in experiment duration using fresh leukopak limited the typical number of voltages we could test to six for each sample. To aid in narrowing the appropriate voltage range, we observed the displacement of the RBCs or rosettes within the sample under bright field microscopy before selecting the fixed voltages and initiating the collection of output samples.

C. Analysis post acoustic separation

Output samples from the side and center outlets along with unprocessed controls were centrifuged at 400g for 5 min. Pelleted cells were treated with 1× RBC lysis buffer (Biolegend) for 15 min at 4C. RBC lysis buffer was quenched with an excess volume of MACS running buffer (Miltenyi Biotec). Cells were enumerated using the Celigo Image Cytometer (Nexcelom). Aliquots of the samples corresponding to 0.5 × 106 cells each were used for surface staining and flow cytometry. Prior to staining, cells were treated with Human TruStain FcX (Biolegend) according to manufacturer instructions. Cells were stained with PE-Cy7 conjugated anti-CD45 (clone 2D1), Pacific Blue conjugated anti-CD3 (clone HIT3a), APC conjugated anti-CD19 (clone 4G7), BV605 conjugated CD56 (clone HCD56), and PE conjugated anti CD14 (clone M4E2). All antibodies were purchased from Biolegend and stained at 1:200 final concentration according to manufacturer’s instructions. Stained cells were washed twice in the MACS running buffer. Dead cells were stained using SYTOX Green (ThermoFisher) at 1:5000 final concentration prior to acquisition in the AttuneNxt Flow cytometer (ThermoFisher). Data were analyzed in FlowJoV10 (BD Biosciences).

D. Recovery and purity calculations

Flow cytometry measured the output purity of CD3+ T cells, CD19+ B cells, CD14+ monocytes, and CD56+ NK cells as a percent of CD45+ leukocytes (see Fig. S1 in the supplementary material for flow gating strategy). Absolute cell numbers from output fractions were calculated from the imaging cytometer results, where we used its measure of the concentration of nucleated cells to quantify the concentration of CD45+ cells counted in the flow cytometry. Recovery at each condition was measured as separation efficiency and defined as the percent of each cell type retained in the side outlet out of the total collected from side and center outlets. For the input samples, comparing the measured PBMC cell concentrations with the concentration of RBCs measured by the hematology analyzer allowed us to calculate the “rosette ratio” (RR), which is the number of RBCs available to bind to a target PBMC.

IV. DISCUSSION

Section:

ChooseTop of pageABSTRACTI. INTRODUCTIONII. MATERIALS AND METHODSIII. RESULTSIV. DISCUSSION <<V. CONCLUSIONSSUPPLEMENTARY MATERIALREFERENCESPrevious sectionNext section

Rosette treatment causes aggregation of unwanted cell types with RBCs. As suspended cells enter the device in the side inlet, they experience acoustic forces that direct them toward the center of the channel. The factors that affect this force are cell volume, density, and compressibility. RBCs, especially because of their higher density, undergo greater acoustic forces than lymphocytes. By binding waste lymphocytes to these RBCs and forming rosettes, the acoustic forces experienced become higher than in any untreated blood cell. The energy required to focus these rosettes to the center is much lower than the amount required to focus lymphocytes; at an intermediary acoustic energy, rosettes containing waste cells can be depleted from the side streams while the unbound target cell type remains. In this work, we demonstrate that this technique is effective in purifying both T cells, which for normal healthy donors are natively high in purity, and B cells, which are natively low.

A. Dependence on RBC concentration

We observed that the resulting output purity of the T cells was limited by the number of RBCs in the leukopak, noting that we were unable to achieve higher than ∼85% purity with native levels of RBCs from a standard leukopak (typically, 1%–3% hematocrit). However, when RBCs from donor matched whole blood were spiked into the leukopak we were able to achieve upwards of 95% pure T cells in the side stream (Fig. 4). To further analyze this result, we normalized for donor variability in input PBMC populations by calculating the total depletion of undesired PBMCs as a fraction of the total undesired PBMCs in the input sample. Figure 7 plots the results of this analysis as a function of hematocrit and shows a positive correlation between them. In other words, we observe the output purity to be a direct function of hematocrit of the leukapheresis product. Since hematocrit levels can be controlled to some extent during leukapheresis collection (by tuning apheresis parameters), we suggest this process could be seamlessly integrated into the current practices for sourcing patient T cells for autologous therapy.

B. Theoretical considerations

The mechanism of the observed purification and its dependence on the number of RBCs available can be further elucidated from a theoretical standpoint. The established theory provides a quantitative description of the acoustophoretic force on a spherical particle in a resonant rectangular microchannel and its resulting trajectory from inlet to outlet.2828. M. Settnes and H. Bruus, “Forces acting on a small particle in an acoustical field in a viscous fluid,” Phys. Rev. E 85(1), 016327 (2012). https://doi.org/10.1103/PhysRevE.85.016327,2929. R. Barnkob et al., “Measuring acoustic energy density in microchannel acoustophoresis using a simple and rapid light-intensity method,” Lab Chip 12(13), 2337–2344 (2012). https://doi.org/10.1039/c2lc40120g As stated in the Introduction, this trajectory will depend on the size, density, and compressibility of the particle, as well as on other parameters of the device. The force on the particle toward the center axis of the channel scales with its volume, V, and acoustic contrast, Φ,where the contrast is a function of the particle's density and compressibility and that of the surrounding fluid. Adopting the simplified approximation of a one-dimensional channel cross section (i.e., infinitely shallow channel) and following Barnkob et al.,2929. R. Barnkob et al., “Measuring acoustic energy density in microchannel acoustophoresis using a simple and rapid light-intensity method,” Lab Chip 12(13), 2337–2344 (2012). https://doi.org/10.1039/c2lc40120g this acoustic force along with the opposing velocity-dependent drag force due to the fluid can be integrated to obtain an expression for the displacement of the particle over time as it flows down the length of the channel in the acoustic field. With the further approximation that the residence time of the particle in the acoustic field is equivalent to the axial position z divided by the average fluid velocity flowing through the channel, a function for the trajectory of the particle in the plane is obtained,where α and β are constants of the acoustic device under fixed operating conditions, y0 is the transverse position of the cell at the inlet, and r is the particle radius. For the rosetted cluster of cells, as opposed to a single spherical particle, we reason that the above analysis can be adapted taking Φ′=…Φ1V1+Φ2V2+Φ3V3…V′,(4)

In other words, V′ indicates the net volume of the cluster as the sum of the volumes of its component cells (indicated by the subscripts 1, 2, 3…), Φ′ is an average of the contrast of the cells that makeup the cluster, weighted by volume, and r′ an effective radius of the cluster. In this analysis, we are mainly concerned with the relative trajectories of an unbound T cell compared with the trajectory of an unwanted lymphocyte bound to some number of RBCs. This simplifies the analysis because many constants of the device, the fluid, the flow rate, the acoustic energy, etc., can be normalized to any convenient value.

Using Eq. (2), with Φ′ substituted for Φ and r′ substituted for r and fixing the constants α and β to achieve trajectories similar to the observed behavior, we arrive at a highly simplified model of the impact of rosettes on the separation of lymphocytes. Figure 8 shows how the model can be used to investigate the significance of the number of RBCs bound to a lymphocyte. Here, we used published values for the acoustic contrast of an average RBC in buffer3030. M. Toubal et al., “Acoustic measurement of compressibility and thermal expansion coefficient of erythrocytes,” Phys. Med. Biol. 44(5), 1277–1287 (1999). https://doi.org/10.1088/0031-9155/44/5/313 and T cell in buffer1717. J. M. Dabbi et al., “Label-free measurement of T-cell activation by microfluidic acoustophoresis,” In 2021 21st International Conference on Solid-State Sensors, Actuators and Microsystems (Transducers) (IEEE, 2021). and diameters 5.6 μm (spherical approximation) and 7 μm, respectively. In the illustrated example, for a lymphocyte entering the channel at its upstream end and 25 μm from the channel sidewall, if unbound or bound to only one RBC, the lymphocyte will be collected at the side outlets. However, with three RBCs bound to it, the undesired lymphocyte is displaced by acoustophoresis to the waste outlet and with 10 RBCs bound reaches nearly its terminal position at the channel center stream.We emphasize that the above calculations and the illustration in Fig. 8 are intended to provide a simplified view of how rosette treatment behaves in acoustophoresis. We have neglected many significant features of the real system including velocity gradients in the fluid flow (Poiseuille flow), three-dimensional non-uniformities in acoustic energy, and distribution of cell properties within each cell type. Additionally, treating the rosette cluster of cells as a sphere of averaged properties neglects effects due to the shape and heterogeneous compressibility of the cluster on the actual acoustophoretic force and effects that are still an area of active theoretical investigation.3131. S. Sepehrirahnama et al., “Acoustic radiation force and radiation torque beyond particles: Effects of nonspherical shape and Willis coupling,” Phys. Rev. E 104(6), 065003 (2021). https://doi.org/10.1103/PhysRevE.104.065003 Despite these simplifications, scaling suggests that dependence on the RBC concentration plotted in Fig. 7 may have room for improvement such that it could require fewer RBCs per depleted cell as the method is further developed, and the antibody cocktail is tailored specifically for acoustophoresis rather than centrifugation.

C. Significance

Current challenges with T cell therapies include labor-intensive processing for cell washing and purification, limited modularity of instruments, and high cost of reagents, including the GMP grade magnetic beads, which are the current standard. This has made handling multiple patients challenging and expensive. Meanwhile, the current methodology of rosette-based separation by centrifuge is effective in isolating desired cells from patient’s whole blood, but this density-based separation requires manual pipet collection of cells from the liquid interface and suffers from slow processing time and operator variability. Furthermore, conventional rosette reagents perform poorly when used on the leukapheresis product rather than whole blood (requiring much higher numbers of RBCs than are available in the leukopak) and, therefore, are not suitable for processing the large number of PBMCs required in the cell therapy. In contrast, integrating the conventional rosetting reagent into our acoustic separation workflow offers several advantages. This approach (1) eliminates the magnetic bead reagent currently used in clinical processing; (2) can accurately modulate recovery and purity by varying the voltage of the transducer; (3) can be integrated into continuous flow systems to enable automation with minimal handling; and (4) involves a closed plastic cartridge that is suitable for high volume manufacture and single use.

Although the results shown here hold promise for future clinical applications, several developments are needed to further advance the technology toward that goal. For practical reasons of sample consumption, these experiments were carried out with the somewhat diluted sample in a single-channel device at a rate of approximately 1 × 106 nucleated cells/min. Elsewhere, we have described a device with 12-channels in parallel that achieved the purification of lymphocytes from untreated leukopak at ∼50 × 106 nucleated cells/min.99. R. Dubay et al., “Scalable high-throughput acoustophoresis in arrayed plastic microchannels,” Biomicrofluidics 13(3), 034105 (2019). https://doi.org/10.1063/1.5096190 In future work, we will test the performance of the rosette approach in the scaled up device and quantify the recovery relative to the input to ensure that cell losses are negligible. In addition to scale-up, we note that the existing commercial rosette reagents are designated for research use only. Therefore, future antibody cocktails would need to meet regulatory requirements along with the process, since incidental infusion of residual antibodies or their components could pose a safety risk. Furthermore, one limitation of this study is that only healthy leukapheresis samples were tested. Additional evaluations are needed with apheresis samples from patients having the relevant diseases since the relative abundance of the mononuclear cell types will often be significantly different and in some cases, diseased cells could exhibit a different acoustic response from their healthy counterparts.

留言 (0)

沒有登入
gif