Research highlights: micro-engineered therapies

Abstract

Lab on a chip systems have often focused on diagnostic, chemical, and cell analysis applications, however, more recently the scale and/or precision of micro-engineered systems has been applied in developing new therapies. In this issue we highlight recent work using microfluidic and micro-engineered systems in therapeutic applications. We discuss two approaches that use microfluidic precision to address challenges in filtering blood – to both remove unwanted pathogens and toxins and isolate rare cells of interest that have therapeutic potential. In both cases chemically-modified surfaces, a bioengineered mannose binding lectin on magnetic particles and antibody-functionalized reversibly degradable alginate film, provide the functionality to remove (or isolate) target cells of interest. The third paper we highlight generates microscale gels as protective niches for cell-based therapies. Importantly, the microgels are designed to have controlled porosity but also mechanical rigidity to protect housed therapeutic cells, like mesenchymal stem cells. We expect continued progress in micro- & nano-enabled therapies facilitated by the fabrication of new microstructured materials, precise separations, and closed-loop sensing and drug delivery.

Dialysis-like therapeutic

Severe microbial pathogen infection in the bloodstream can lead to sepsis, a full-body inflammatory response that potentially causes septic shock, multi-organ failure, and death.^1,3 Current treatment for sepsis includes fluid resuscitation and broad spectrum antibiotics. Following identification of the infection source and antibiotic susceptibility, targeted antibiotics can be more effective and reduce side-effects.² Kang et al. propose a new class of therapy – an extracorporeal blood cleansing system to remove pathogens as well as their endotoxins without the initial need to identify the source of infection.³

The blood cleansing system is comprised of two key components: (i) a genetically engineered mannose binding lectin (MBL) protein and (ii) a bio-inspired microfluidic spleen acting to remove functionalized magnetic particles from whole blood. The carbohydrate-recognition domain and neck domains of human-MBL were fused with IgG1 Fc, forming Fc-containing MBL (FcMBL) shown in Fig. 1a. FcMBL also greatly decreased coagulation-promoting, complement-activation, and DNA-binding activities compared with native MBL. The FcMBL was then biotinylated and uniformly coated on streptavidin-coated superparamagnetic nanobeads to create multivalent magnetic opsonins (Fig. 1b). These magnetic opsonins are then added to septic blood and thoroughly mixed. In order to remove the magnetically opsonized pathogens and endotoxins, a branched microfluidic device was used. Specifically, the channel consisted of two parallel rectangular channels with slits joining the two, allowing for bead cross migration under applied magnetic force. One channel contains sterile, isotonic saline and the other whole blood with magnetic particles that are then drawn into the saline solution by an induced magnetic field gradient (Fig. 1c).

Fig. 1 Microfluidic blood cleansing. A. Schematic of genetically engineered Fc-containing MBL. B. Functionalized magnetic beads binding to pathogens. C. Schematic of the blood-cleansing system that mimics the spleen functionality. D. Experimental animal setup for extracorporeal blood cleansing. E. and F. In vitro studies of E. coli and LPS levels after treatment, respectively. G. and H. In vivo studies of S. aureus and E. coli levels after treatment (P < 0.01 for S. aureus and P < 0.004 for E. coli). I. Survival and J. LPS intensity after a lethal dosage of LPS was injected after treatment. All images were adapted from Kang et al. with permission.³

Human blood was spiked with cecal contents of meat-fed rats (3.5 × 10⁴ anaerobe CFU and 1.5 × 10⁴ aerobe CFU in 10 mL) in an anaerobic environment to mimic intra-abdominal sepsis. After a single pass through the device, >98% and >80% of anaerobic and aerobic bacteria were removed using a flow rate of 10 mL h⁻¹. Using fresh whole human blood spiked with S. aureus (10⁴ CFU mL⁻¹) about 60% of the bacterial load was removed with each pass. Decreasing E. coli levels in spiked blood are shown in Fig. 1e. Additionally when fresh whole human blood was spiked with LPS endotoxin (10 μg ml⁻¹) and continuously flowed through the device, endotoxin levels decreased by a factor of 4 (Fig. 1f). Lastly, because only ~80% of unbound single nanomagnetic beads could be recovered due to their small magnetic moment, larger 1 μm uncoated magnetic beads were added along with the 128 nm FcMBL particles to create stronger local field gradients to attract smaller particles and generate a larger collective magnetic moment. This increased magnetic capture to 99.6 ± 0.6% under the same flow conditions.

Similar to the in vitro studies, the device in vivo removed approximately 90% of live S. aureus and E. coli within 1 h, experimental setup shown in Fig. 1d. Significantly lower levels of pathogens were found in septic rat circulation following blood cleansing compared to uncleansed controls or those treated with the biospleen without added magnetic opsonin beads or using beads lacking the FcMBL coating (Fig. 1g and h), showing the importance of FcMBL in effective blood cleansing. Finally, rats were injected with a lethal dose of LPS endotoxin and then treated with the biospleen device. Treatment resulted in a significant decrease in LPS endotoxin (Fig. 1j). Although 86% of untreated rats died after LPS injection, 89% of rats treated with the biospleen device survived for the entire 5 h experiment (Fig. 1i).

In addition to showing promise for initial sepsis treatment by way of pathogen and endotoxin removal, this extracorporeal blood cleansing device may also be used to collect bacterial cells and identify the source of infection. Therefore, this tool can be coupled with existing methods to further eradicate specifically targeted microorganisms.

Microfluidic capture of therapeutic cells from blood

Besides separations aimed at removing pathogenic cell populations, approaches to isolate and concentrate cells from body fluids that have therapeutic use can benefit from microfluidic technology. It has been shown that endothelial colony-forming cells (ECFCs) have great potential for therapeutic purposes such as vascular therapies.⁴ The exact origin of ECFCs is unknown, but capturing these cells from peripheral blood in a way that can be used for clinical applications remains a need. Cytometric methods may cause damage to cells, however microfluidic approaches can solve that problem for high throughput applications.

Lin et al. have introduced a microfluidic platform for this purpose that has an antibody-laden hydrogel coating for capturing CD34⁺ cells.⁵ Micropost surface modification (anti-human CD34 antibody functionalized posts) was carried out as described by Hatch et al. using an alginate gel.⁶ Then whole blood (6 mL) was flowed through 20 parallel microfluidic devices at 5 μL min⁻¹ (Fig. 2A). The blood was rinsed from the device by MES buffer followed by injection of ethylene diamine tetraacetic acid (EDTA) in PBS to chelate Ca²⁺, dissolve the alginate gel and elute cells for downstream culture. In addition, another 6 mL whole blood was processed using standard isolation methods to be compared with microfluidic-separated cells.

Fig. 2 Microfluidic platform for capturing endothelial colony-forming cells (ECFCs). A. 6 mL blood was processed with the microfluidic device (mf-ECFCs), and 6 mL was processed using a standard centrifugal method (std-ECFCs). B. Phase contrast image of an ECFC colony captured by microfluidic device and total number of captured colonies by both methods. (Scale bar: 500 μm). C. Representative image and quantification of mf-ECFCs for in vitro assessment of ECFC function. Number of bound leukocytes increased after TNF-α treatment (scale bar: 200 μm). D. Capillary-like networks assembled by mf-ECFCs on Matrigel coated plates. Green fluorescence stain shows viable cells, and red shows dead cells. Total tube length per unit of area was quantified after 24 h. (Scale bar: 500 μm). E. mf-ECFCs and mesenchymal stem cells were embedded in collagen–fibrin gel and the mixture was injected into nude mice. Macroscopic view of the site of implantation after 7 days is shown. (Scale bar: 2 mm). F. Microvessel density at day 7 for both groups was quantified. All images were obtained from Lin et al. with permission.⁵

Processing time for both approaches is about 60 minutes. The microfluidic approach yielded statistically similar number of ECFCs compared to the standard method (Fig. 2B). However, the standard method requires several initial steps such as density gradient centrifugation, erythrocyte lysis and so on to obtain the erythrocyte-free fraction.^4,7 The microfluidic approach has the advantage of running unprocessed blood, which enables the possibility of recirculating the CD34-depleted blood fraction back into the patient, and therefore facilitate processing larger volumes with reduced effect on patients.

Because of the functional differences between the various sub-populations of endothelial progenitor cells, performing distinctive phenotypic characterizations is important. In this study the authors compared microfluidic-ECFCs with standard-ECFCs from the same volunteer subjects in different ways such as the capacity to form vascular networks in vivo (Fig. 2E and F), and the results show that they are highly similar in molecular phenotype and function. Microfluidic-ECFCs uniformly expressed endothelial cell markers (CD31, vWF, and VE-cadherin), but not hematopoietic (CD45) and mesenchymal (CD90) markers. Isolated cells also showed endothelial functions: (1) binding of UEA-1 lectin; (2) in vitro capillary-like network formation ability (Fig. 2D); (3) proliferative and migratory responses to angiogenic growth factors; (4) upregulation of leukocyte adhesion molecules upon exposure to an inflammatory cytokine (TNF-α) and increase in leukocyte binding (Fig. 2C).

In summary, this new microfluidic platform facilitates processing blood for therapeutic applications, and the results validated functional and phenotypical properties of captured circulating ECFCs from peripheral blood.

Engineered injectable microscale cell niches for cell therapy

Once therapeutic cells are isolated, delivery is often a subsequent challenge. Cell therapy efficacy suffers due to limitations in the strategies used for cell delivery into hosts. Injection of cells free in suspensions has been shown to result in nonspecific cell incorporation, and injections of pre-formed aggregates both require high doses and risk mechanical injury to the cells. Approaches involving responsive biomaterials that are co-injected with cells and gel in situ are an improvement but do not allow pre-priming of the gel environment by indwelling cells, which is necessary for efficient therapeutic response in damaged tissue. To address these problems, Li et al. developed injectable 3D microscale cell niches (microniches) constructed from biodegradable gelatin microcryogels to serve as delivery vehicles.⁸ These microniches are formed in vitro where seeded cells can be primed to increase extracellular matrix (ECM) deposition and cell–cell interactions to more closely resemble tissue and are then injected into the lesion. In their work, the authors demonstrate the delivery efficacy of the system in a critical limb ischemia (CLI) mouse model.

The gelatin microcryogels (GMs) are formed via cryogelation of gelatin precursor solution within a PMMA microstencil chip containing arrays of circular wells. The resulting network of GMs has high flexibility and compressibility allowing for easy injection and corresponding high porosity (~95%) facilitating automatic cell loading via absorption. Adipose-derived mesenchymal stem cells (MSCs) were seeded into the harvested GM network at densities ranging from 100 to 1200 cells per GM. Cell–cell interactions and ECM deposition by cells within the GM networks was optimized by controlling the cell densities and the duration of priming (culture time post-loading into the GMs) It was shown that priming for 2 days with high cell density was optimal (Fig. 3). Secretion of angiogenic factors by the embedded MSCs also increased under the same optimized priming conditions. Once the cells were loaded and allowed enough time to deposit ECMs and form intercellular contacts, the resulting microniches were washed, strained, and injected into the mouse model.

Fig. 3 Cell priming in GMs to constitute the 3D microniches. (A) Two experiments for priming hMSCs in GMs. (B) SEM images of ECMs of hMSCs in GMs under the two experimental conditions. (C) Gene expression of major ECM molecules from hMSCs in GMs under the two experimental conditions. (D) Protein expression of representative ECMs and E-cadherin for hMSCs in the two experimental conditions. *P < 0.05; **P < 0.01 (n = 3). Image was obtained from Li et al. with permission.⁸

The authors compared therapeutic benefits of microniche-based delivery to free-cell injections. Luminescent signal readings of firefly luciferase-positive MSCs were used to show that microniche-based injections yielded better overall retention and survival of cells than free-cell injections with local retention lasting up to 14 days in the microniche group but diminishing after 4 days in the free-injection group. Using the CLI mouse model, the authors showed that while the mice in the free-cell injection group suffered 80% limb loss after 28 days, the microniche group showed limb salvage in 6 out of 8 mice, 4 of which showed complete recovery. Microvessel and arteriole formation and blood perfusion was also enhanced in the microniche group compared to the free-injection group while muscle degeneration was reduced. Importantly, the number of MSCs seeded into each microniche (10⁵) was an order of magnitude less than the number of cells in each free-injection (10⁶).

The results indicate that the priming of the cells enabled by proper encapsulation is an important step both for ensuring robust delivery and for efficient incorporation into the in vivo environment and serve as a convincing demonstration of such a delivery micro-engineered vehicle.

References

1. T. Glück and S. M. Opal, Advances in sepsis therapy, Drugs, 2004, 64(8), 837–859.
2. J. Garnacho-Montero, et al., Impact of adequate empirical antibiotic therapy on the outcome of patients admitted to the intensive care unit with sepsis, Crit. Care Med., 2003, 31(12), 2742–2751.
3. J. H. Kang, et al., An extracorporeal blood-cleansing device for sepsis therapy, Nat. Med., 2014, 20, 1211–1216.
4. J. Melero-Martin, Z. A. Khan and A. Picard, et al., In vivo vasculogenic potential of human blood-derived endothelial progenitor cells, Blood, 2007, 109, 4761.
5. R. Z. Lin, A. Hatch, V. G. Antontsev, S. K. Murthy and J. M. Melero-Martin, Microfluidic capture of endothelial colony-forming cells from human adult peripheral blood: phenotypic and functional validation in vivo, Tissue Eng., Part C, 2014 DOI:10.1089/ten.tec.2014.0323.
6. A. Hatch, G. Hansmann and S. K. Murthy, Engineered alginate hydrogels for effective microfluidic capture and release of endothelial progenitor cells from whole blood, Langmuir, 2011, 27, 4257.
7. Y. Lin, D. J. Weisdorf, A. Solovey and R. P. Hebbel, Origins of circulating endothelial cells and endothelial outgrowth from blood, J. Clin. Invest., 2000, 105, 71.
8. Y. Li, et al., Primed 3D injectable microniches enabling low-dosage cell therapy for critical limb ischemia, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 13511–13516.

---