ChemE696.ProjectO

From MicrofluidicsWiki

Title: Microfluidic Devices for the Study of Gas Embolotherapy

Author(s): Robinson Seda

Date Submitted: April 10, 2009

Multimedia Presentation: click here

Summary

Hepatocellular carcinoma (HCC) is the most common form of liver cancer and one of the most common tumors worldwide. Liver transplantation is the current choice to treat this malignancy, but many patients die while on the wait list. Embolotherapy (occlusion of blood vessels for therapy) using solid emboli has succeeded in extending the patient’s lifespan until a liver donor is available, but restricting its delivery can be difficult. The proposed technique involves bubbles that originate from microdroplets using high intensity ultrasound that will eventually lodge the capillaries “starving” the tumor to death. Microfabricated capillary networks have been developed with (poly)dimethylsiloxane (PDMS) using soft lithography and PDMS casting techniques to simulate the tumor microcirculation. These devices have been used to understand the different aspects of bubble delivery and lodging, but one extension to the current research will be the endothelialization of such networks providing a more biologically compatible environment to develop this intriguing and revolutionary treatment.

Introduction & Motivation

Figure 1: Concept of gas embolotherapy using Acoustic Droplet Vaporization

Hepatocellular carcinoma (HCC) is the most common form of liver cancer and one of the most common tumors worldwide. It can have three different growth patterns that include a single tumor, multiple or a poorly defined tumor with an infiltrative growth pattern. Liver transplantation and surgical resection are the most common ways to treat this malignancy, but unfortunately surgery is only possible in 10-15% of the cases and also the patient may die while on the waitlist for a liver donor [1]. Embolotherapy or embolization has had special attention in the treatment of HCC because it can be used to treat unresectable tumors and has shown success in extending the patient’s lifespan up to 2 years (depending on the patient’s medical condition) [2]. This procedure involves passing a catheter percutaneously through the femoral artery then to the abdominal aorta and finally through the common hepatic artery until the branch supplying the tumor is detected. Smaller catheters are passed through until reaching the feeding vessel of the tumor and once there, small particles (emboli) and drugs are injected to block the blood flow to the tumor. However, controlling the delivery of the emboli to the tumor can be very difficult and also healthy tissue may be at risk of ischemia.

A novel technique developed by the Biotransport Group will involve the occlusion of blood vessels using bubbles that originate from albumin-coated perflurocarbon (PFC) droplets. These droplets will be injected intravenously into the bloodstream and vaporized using focused high intensity ultra sound, technique that was named Acoustic Droplet Vaporization (ADV). The resulting gas bubbles (~120 times larger than the microdroplets) [3] will be large enough to lodge in the blood vessel stopping the flow and starving the tumor to death (see Figure 1). This technique allows selective delivery of the embolus to the tumor and is also minimally invasive as opposed to conventional embolization using solid emboli.

In order to develop gas embolotherapy, several aspects of bubble dynamics have to be investigated including bubble vaporization, bubble transport and lodging [4-9]. Microfabricated capillary networks have been developed using (poly)dimethylsiloxane (PDMS) to simulate the tumor microcirculation. The bubble, which will be transported by blood flow, is expected to lodge, but the conditions at which this lodging will occur need to be understood. Effects of bubble size and vessel diameter are examples of the focus of previous research [9], while the study of the bioeffects to which blood vessels and healthy tissue are exposed will be addressed by future investigations.

Experiments on Bubble Lodging

Calderon et al., 2006

Figure 3: Schematic of experimental setup

Figure 2: PDMS Microchannels with single and multiple bifurcations.

Device Description

The capillary networks used in these studies were developed following standard PDMS casting techniques described by [9]. This technique involves the casting of PDMS using an SU-8 mold and bonding the resultant pattern to a slab of unpatterned PDMS with oxygen plasma. The base to curing agent ratio used was 10:1, which provides an elastic modulus comparable to physiological values for microvessels [11]. These microchannels consisted of a single or multiple bifurcations and ranged from ~80 to 100 microns in diameter. Cross sections included D-shape, circular and square, while other parameters such as bifurcating angle (angle between daughter channels), perimeter ratio and hydraulic diameter ratio were also investigated. Two reservoirs were attached at the inlet and outlet of the microchannel to induce flow.

Device Operation

Deionized water is pushed through the channels by a driving pressure (∆p) specified by the two reservoirs at different heights (∆h). A microsyringe (26-gauge needle, 10 µL Hamilton syringe) was used to inject an air bubble into the channel and the ∆h was adjusted until the bubble lodged. This was known as the lodging pressure. Once lodged, the pressure difference was increased until bubble transport was resumed. This was known as the dislodging pressure.

Results

Figure 4: Lodging and dislodging pressures for all cross sections (A-E). Data for which p<0.05 is indicated by * (from Calderon et al., 2006) .

Lodging and dislodging pressures

The obtained results suggested that the dislodging pressures will be higher than lodging pressures for all different cross sections. Higher pressures were obtained in microchannels with smaller perimeter ratios (parent to daughter channel). More importantly, these dislodging pressures lied below physiological values implying that once the bubbles are lodged, they will remain lodged in the vasculature [10].

Bubble splitting and reversal

Bubble splitting and reversal at the bifurcations was also reported. Once the bubbles reached the bifurcation, two states of bubble lodging were observed (see Figure 5 (a) and (b)). While most of the bubbles lodged in the first state, they also experienced the second lodging state when the pressure was increased, thus lodging at a higher pressure. This behavior was more commonly observed in microchannels with a larger bifurcating angle. Increasing the pressure during the second state induced bubble reversal to one of the daughter branches shunting the flow to the unoccluded branch (see Figure 5 (c)) . After this, no matter how much the pressure was increased, the bubbles did not dislodge and the flow was blocked. These results were confirmed by transporting microspheres into the channels.

Figure 5: Stages of bubble lodging (a), splitting (b) and reversal(c) at the bifurcation (from Calderon et al. 2006) .

Multiple bifurcation model

A second set of experiments was carried out using a multiple bifurcation network. This network was successfully occluded using several bubbles, which in turn needed even higher dislodging pressures compared to the single bifurcation model. It is expected that by introducing several bubbles into the tumor microcirculation, the flow will be completely stopped.

Theoretical Model

Figure 6: Theoretical curves and experimental data for non dimensionalised pressure (from Calderon et al. 2006) .

In order to predict bubble lodging a theoretical model was proposed by [11] based on non dimensionalised pressure. Applying the Young-Laplace equation at the bubble interfaces we get:

and

Where the radius of the capillary tube was replaced as an approximation by half of the hydraulic diameter of the parent (HD) and daughter (Hd) channels and β and α being the contact angles of the parent and daughter channels, respectively. Combining these two equations we get an expression that relates the front (P_i) and rear (P_o) menisci pressures:

Rearranging the equation an introducing the non dimensional parameter , the equation becomes:

and in non dimensional form

This equation was plotted against different values of at different contact angles and compared to the experimental results, which seem to be in agreement with the model (see Figure 6).

Experiments on Bubble Injury

Wong et al., 2006

Every blood vessel of our body is lined with a thin layer of endothelium that, among other functions, reduces friction during the transport of blood. This layer may play an important role in bubble transport through interactions at the interface and change in surface roughness or lubrication. Previous work has showed the potential of coating PDMS channels with human endothelial cells [10, 11], but bubble injury in microvessels has received little attention. A recent investigation from the Biotransport Group [11] showed the potential of bubbles to injure the endothelium layer by quantifying cell viability, but the mechanisms of injury are yet to be understood.

Device Description

Fabrication

Figure 7: Microchannel fabrication using non conventional PDMS casting techniques (Wong et al., 2006)

For these experiments a non conventional casting technique was used as described by Wong et al., 2006. A thin wire (~100microns) was used as a positive mold to cast PDMS and create a circular cross section. A device was machined, which consisted of a base and a mold to hold the wire in place (see Figure 4). The wire was passed through holes in the mold while two knobs at the ends of the base were tighten to keep the wire straight. The PDMS was poured in the mold holding the wire in place and after curing, the wire was pulled leaving a circular lumen in the transparent piece of PDMS.

Endothelialization

Figure 8: Fluorescence microscopy images of endothelial cells at different sections of the channel (from Wong et al. 2006) .

To create a biologically compatible setting for experiments these devices were coated with a monolayer of human endothelial cells. This was achieved by flowing fibronectin through the channel to create a "primer" prior to the seeding of the cells as described by [10, 11]. Cells from passages 3-11 were cultured in an incubator at 37 degrees Celsius and 5% of CO₂. After the cells attached to the surface, they were monitored to determine cell viability using a fluorescent dye (calcein AM)that will show up green (live cells) under fluorescence microscopy (see Figure 8). More than 90% were viable cells [11].

Bubble Injury

Air bubbles of different sizes were pushed through the endothelialized channel at different flow rates using a syringe pump. Cell viability was determined using fluorescence microscopy (EthD-1) and morphology studies to look for dead cells. Dead or injured cells will absorb the fluorescent dye and turn read (see Figure 9).

Figure 9: Fluorescence image of endothelial cells dyed red indicating injury to the cells following the passage of a long bubble (from Wong 2009)

Extension to Current Understanding

Previous work has demonstrated the potential of these microchannels to simulate the microcirculation. They have also showed different conditions at which bubble lodging may occur in the microvasculature by developing a theoretical model based channel geometry and contact angles. However, these experiments did not consider the endothelial cell layer. Bubble injury was also investigated showing the likelihood of a bubble to injure endothelial cells during bubble transport. However, the mechanisms under which this injury occurs are yet to be understood. Future experiments will focus on the quantification of stresses and stress gradients that may cause bubble injury and also the effects of the glycocalyx layer of the endothelium on bubble lodging in bifurcation channels. Understanding the set of conditions that lead to cell damage by bubbles will be necessary as preventive means to further develop and improve the gas embolotherapy technique.

Conclusion

Microfabricated capillary networks have been developed with (poly)dimethylsiloxane (PDMS) using soft lithography and PDMS casting techniques to simulate the tumor microcirculation. These devices have been used to understand the different aspects of bubble vaporization, transport and lodging, yet other questions remain. Understanding the mechanisms by which bubbles affect cell viability and how cells might change the conditions just described will be key to develop anti-injury strategies leading to a further step towards the development of gas embolotherapy.

Acknowledgements

Thanks to the Biotransport Research Lab, specially Zheng Zheng Wong for providing information on device fabrication and microchannel endothelialization.

References

1. Sean O. Stitham, MD; James R. Mason, MD; Hepatocellular carcinoma, MedlinePlus, http://www.nlm.nih.gov/medlineplus/ency/article/000280.htm, (2008)
2. Chemoembolization, RadiologyInfo, http://www.radiologyinfo.org/en/info.cfm?PG=chemoembol, (2008)
3. Kripfgans, Oliver D. et al.; Acoustic Droplet Vaporization for therapeutic and diagnostic applications; Ultrasound in Med. & Biol., Vol. 26, No. 7, pp. 1177–1189, (2000)
4. Bull, J.; Cardiovascular Bubble Dynamics; Critical Reviews in Biomedical Engineering, 33(4):299–346 (2005)
5. Calderon, A. J.; J. B. Fowlkes; J. L. Bull; Bubble splitting in bifurcating tubes: a model study of cardiovascular gas emboli transport; J Appl Physiol 99: pp. 479–487, (2005)
6. Eshpuniyani, B.; J. B. Fowlkes; J. L. Bull; A bench top experimental model in bubble transport in multiple arteriole bifurcations; International Journal of Heat and Fluid Flow 26: pp. 865–872, (2005)
7. Ye, T.; J. L. Bull, Direct Numerical Simulations of Micro-Bubble Expansion in Gas Embolotherapy; Journal of Biomechanical Engineering, Vol. 126, pp. 745-759, (2004)
8. Ye, T.; J. L. Bull; Microbbuble Expansion in a Flexible Tube; Journal of Biomechanical Engineering, Vol. 128, pp. 554-563, (2006)
9. Calderon, A. J., et al., Microfluidic Model of Bubble Lodging in Microvessel Bifurcations, Applied Physics Letters, 89 (2006)
10. Shi, M. et al., Endothelialized Network with a Vascular Geometry in Microfabricated (Poly)dimethyl siloaxane, Biomedical Microdevices 6:4, pp. 269–278, (2004)
11. Wong, Z. Z.; Stephen, J.; Bull, J., Effects of bubbles on cell viability in a circular-lumen endothelialized microvascular model, America Society of Artificial Internal Organs 52(4), p494, (2006)
12. Wong, Z. Z.; Gas Embolotherapy: Bubble Evolution in Acoustic Droplet Vaporization and Design of a Benchtop Microvascular Model; University of Michigan, PhD Dissertation, (2009)