A Microfluidic Platform for the Investigation of Transport ... · A Microfluidic Platform for the Investigation of Transport in Small Blood Vessels ... testing and validation of a
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
A Microfluidic Platform for the Investigation of Transport in Small Blood Vessels
by
Sascha Pinto
A thesis submitted in conformity with the requirements for the degree of Master of Applied Science
Institute of Biomaterials and Biomedical Engineering
A Microfluidic Platform for the Investigation of Transport in Small Blood Vessels
Sascha Pinto
Master of Applied Science
Institute of Biomaterials and Biomedical Engineering University of Toronto
2012
Abstract
The microvasculature has the main function of transport of dissolved gases, nutrients and waste
between blood and tissue. Systematically probing transvascular transport rates in these vessels under
well defined conditions is challenging. In vivo and in vitro studies are characterized, respectively, by
limited optical access and control over perfusion concentrations and failure to resemble the structure
and function of an intact organ. In this thesis, I present the development of a microfluidic platform
for investigating molecular transport across mouse mesenteric arteries (150-300μm diameter) in a
controlled physico-chemical microenvironment. Intact vessels are perfused with 4 kDa FITC-
Dextran and the permeation coefficient of this molecule across the vessel wall is quantified using
laser scanning confocal microscopy paired with a 2-D numerical model. Functional viability of the
examined vessel, through phenylephrine and acetylcholine dose responses, is probed, and shear and
phototoxic effects are reported.
iii
Acknowledgments
First and foremost, I would like to thank my parents and my brother for their continued support
throughout this project. Knowing that I could always rely on them (mentally, financially, and
nutrionally) made this project possible.
I would like to acknowledge the help and vast knowledge of Sanjesh Yasotharan, my project brother
throughout these last 2.5 years. From fabrication to soldering, thanks for being a colleague and
teacher.
I would like to thank the members of the Bolz lab, my family away from home. No matter the
frustrations, experiments were simply fun to do with you all.
I would also like to thank the members of the Günther lab. I hope you keep the Starbucks and after
school traditions going.
In addition, I would like to thank my supervisor Axel Günther for his continued support, discussion
and impartment of knowledge, and understanding when I felt homesick. Also, for the timely Ritter
Sports…but only the yellow ones.
I would also like to acknowledge the supervision of our collaborator Steffen-Sebastian Bolz (I will
always know that the hardest part is always ahead), as well as the advice and experimental guidance
from our collaborators Dr. Dan Dumont and Paul Van Slyke, and insight from my committee
members Craig Simmons and Dr. Myron Cybulsky.
Many thanks goes to my close friends back home for their continued understanding and support
these last few months.
Throughout the course of this project, I was funded by NSERC CGS M and FQRNT B1.
Finally, as Meghan has taught me, I would like to thank the mice. There would be no data or results
without them.
iv
Table of Contents
Acknowledgments .......................................................................................................................................... iii
Table of Contents ........................................................................................................................................... iv
List of Figures ................................................................................................................................................. vi
List of Appendices ........................................................................................................................................ vii
List of Abbreviations ................................................................................................................................... viii
1.2 Specific Aims ......................................................................................................................................... 3
1.3 Background: The Microvasculature and the Endothelium ............................................................ 3
1.4 Previous Work on Transport in the Microvasculature ................................................................... 8
2.5 Numerical Model ................................................................................................................................ 20
A.2 Post-Processing Script for Confocal Images ........................................................................................ 42
A.3 Numerical Model ...................................................................................................................................... 47
A.4 VEGF as a Positive Control ................................................................................................................... 49
A.5 Local Transport Study .............................................................................................................................. 51
measurements were complemented with dose-response measurements that assess the functional
viability of the artery endothelial and smooth muscle layers
An important feature of our approach is the capacity for controlled perfusion through the vessel’s
lumen. Shear effects on the endothelium have been broadly investigated.61-65 High shear stresses
applied on the endothelium are known to induce the production of endothelial nitric oxide species
(eNOS), activating the dilation mechanism in arterioles and venules and increasing vascular
permeability.65-67 Vessels obtained from the mouse mesentery bed have been previously shown to
produce eNOS species when encountering wall stresses greater than 10 dynes/cm2.68 With the range
of diameters used throughout these experiments 10, a perfusing flow rate of 5 μl/min was used
throughout the experiments, resulting in approximately 1.06 dynes/cm2 in a perfused 200 μm
diameter vessel.68 Figure 3.1a indicates that the functioning of the SMCs is not affected by luminal
flow, as the constrictory mechanism is preserved in the presence of the applied luminal flow. The
unaffected constrictions suggest that there are no dilatory forces present due to the applied flow to
counteract the constrictory effect of the superfused PE. The presence of a dilatory response to ACh
(Fig. 3.1b) demonstrates functioning of the endothelium in the presence of shear. Dilatory
measurements carried out in the presence of shear show a weakened dilation at high concentrations
30
of ACh. Resistance arteries have been previously shown to have constrictory responses in the
presence of high (>10-6 M) concentration of ACh.69, 70
Figures 3.2-3 provide the experimental results obtained during laser scanning confocal microscopic
measurements taken on the fixated vessels. Figure 3.2b illustrates the autofluorescence of the fixated
vessel at the voltage gain used for measurement. The location of the endothelial barrier is also seen,
utilizing vessels obtained from a transgenic mouse with a GFP tagged to an endothelial specific Tie2
receptor. Figure 3.2c illustrates the transient perfusion of the fluorescent marker through the vessel,
after which, the marker is seen to permeate transvascularly and after some time, the intensity profile
stabilizes (Fig. 3.2d, 3.3a). The transient change in permeation, seen for one vessel in Fig. 3.3a,
shows an early, abrupt rise in the marker’s abluminal concentration, followed by a steadying
interaction with the convective microenvironment. The change in average intensity from frame to
frame is then seen to decrease to zero. The variations in the intensity measurements at steady-state
(Fig. 3.3a) is owed to the fixed pattern noise in the confocal measurement signal due to the high
pixel scan times 71, as well as movement of the vessel due to slight pressure variations in the
perfusion and superfusion flow channels. The steady-state concentration profiles (Fig. 3.3b) show a
linear decrease within close proximity of the vessel wall, the length of which is defined to be the
concentration boundary layer.72
The linear relationship obtained from the numerical model is a result of the transport mechanisms
present, and how they are affected by the vascular wall’s permeability coefficient. After permeation,
a luminal concentration will be present at the interface with the vascular wall, a value of which will
be determined by the inlet concentration and luminal flow rate. Transport through the vascular wall,
a summation of both transcellular and paracellular transport, is assumed to be primarily diffusive for
the molecule used in these experiments (Stokes’ radii of approximately 1.4 nm).50 The concentration
gradient present in the direction of diffusive transport is linear. Therefore, the value of P will
31
determine the concentration present at the interface of the vascular wall with the abluminal
microvironment. Abluminally, the concentration gradient within the concentration boundary layer
will be linear as well.72 Therefore, as P changes in the vascular wall, a linear change results in the
abluminal concentration gradient, as seen in the relationship in Fig. 3.4b..
Figures 3.5-6 summarize the permeability and viability results from the experimental data sets. As
previously mentioned, the first excitation and intensity measurement method focused on the whole
vessel. The viability data (Fig. 3.5a-b) from these experiments saw a weakening of the constrictory
and dilatory response after each perfusion and excitation of 4kDa FITC-Dextran. The weakening of
the vessels endothelium and SMC layer resulted in continually increasing calculated permeability
coefficients (Fig. 3.5c-d). The prolonged laser scanning times (30 sec. intervals, every 1 min. for
approximately 40 mins. for each perfusion of FITC-Dextran) used through the experiment are
thought to have produced free oxygen radicals, damaging the ECs and SMCs.54 The protocol was
therefore modified, so as to have controlled excitation and analysis, as described in section 2.3. With
excitation of the fluorophore outside of the vessel wall, less phototoxicity occurred, as seen in the
PE and ACh dose responses (Fig. 6a-b). There was no decrease in constrictory responses after
subsequent perfusions of FITC-Dextran, and the vessel showed a strong dilatory response to
increasing doses of ACh, ensuring a healthy and functionally intact endothelial layer. The calculated
permeability coefficients reflected the maintained viability, as the obtained values are consistent
throughout sequential perfusions. The average P obtained from controlled excitation was 2.04x10-6
cm/s (Fig. 3.6) with lower and upper limit measured values of 1.437x10-6 cm/s and 3.334x10-6 cm/s
respectively. Errors affecting the measurement of this value include: mesh dependence used in the
numerical model (Appendix A.3), vessel wall size variation affecting representation of numerical
model parameters, vessel wall size affecting abluminal flow profile (effects of abluminal Pe variation
see in Appendix A.3), vessel wall non-uniformity affecting abluminal concentration profile, precision
32
of the linear-fit in post-processing, and spatial and temporal variations within confocal
measurements.
33
A microfluidic platform was developed with the ability to reproducibly quantify the permeability
coefficient of a mouse mesenteric artery to 4 kDa FITC-Dextran. The ability to abluminally deliver
vasoactive substances throughout the perfusion of the fluorescent marker allowed for probing of the
isolated artery segment’s viability and health of the endothelium, the principle barrier to
transvascular transport. Flow was applied through the lumen in a controlled manner and the effect
of the applied shear on the vessel’s viability was quantified. A method for exciting and measuring the
perfused and permeating fluorescent marker was developed so as to reduce phototoxicity and
maintain the integrity of the transport barrier. A post-processing script was developed to analyze
confocal measurements and obtain spatially and temporally averaged steady-state concentration
profiles of the permeated fluorescent marker. Profiles were paired with a numerical model that was
solved for different permeability coefficients. The experimental permeability coefficient was
obtained through backwards fitting the experimental concentration profiles to the solution obtained
from the numerical model. The method obtained permeability coefficients in an accurate manner
(7.06% deviation, n=4) and allowed for viability probing of the whole artery segment in a controlled
microenvironment. This enabled relating of the phototoxic effects produced from whole vessel
illumination to the vessel’s viability and measured permeability coefficient.
34
A proposed strategy to further quantify the developed method is to examine the effect of a
substance known to alter the permeability of the vessel being studied. Vascular endothelial growth
factor (VEGF) has been previously shown to induce hyperpermeability in the microvasculature
through a variety of mechanisms.73-76 Preliminary experiments examining the effect of VEGF on the
constrictory and dilatory responses of a fixated mouse mesenteric artery were carried out (Appendix
A.4), and showed an irreversible weakening effect on the constrictory and dilatory responses,
indicating a biological response. Preliminary permeability experiments examining the effect of
VEGF however showed a negligible increase (Fig. A.4). A revised protocol for examining the effect
of VEGF or other permeability inducing substances (eg. thrombin, histamine and bradykinin 77, 78) is
required.
Permeability is seen to vary in similar sized vessels across different vascular beds. Arterioles from the
brain microvasculature are characterized by an endothelium with a higher prevalence of tight
junctions, resulting in decreased permeability coefficients across the blood brain barrier (BBB).4, 38
Further investigation into the BBB could be performed with the presented method, albeit, with a
modified microfluidic design and fluidic delivery method capable of lower flow rates. Our group has
developed a platform capable of fixating and studying the viability of mice cerebral olfactory
arteries.79 With this platform serving as the groundwork, and with a modified protocol allowing for
perfusion and ACh measurements of the artery segment, permeability coefficients for cerebral
vessels could be obtained.
35
In vivo investigation of local transport mechanisms is currently limited by surrounding tissue and
imaging capabilities. With a microfluidic platform designed on a coverslip glass substrate, the
working distance of the current method (~2 mm) would decrease, increasing magnification and
numerical aperture for cell-level imaging. Paired with a transgenic mouse, which would allow for
localization of the endothelial barrier (Appendix A.5), and imaging techniques that can measure
transport dynamics, such as fluorescence recovery after photobleaching (FRAP)80, 81, local transport
rates and mechanisms can be investigated.
As previously mentioned, a main feature of this platform is the ability to apply controlled perfusion
flow rates. This, with the ability to probe viability in the presence of a luminal flow can make the
platform an ideal drug development assay. The micro-scale fluidic channel dimensions allow for low
reagent use, while controlled superfusion and perfusion flows can administer substances of interest
in a temporal and spatial 23 fashion. Measurement of the permeability of a certain vessel to a specific
drug can improve pharmacokinetics, currently the main cause of failure in drug development.
36
1. R. K. Jain, Cancer and Metastasis Review, 1987, 6, 559-593.
2. Y. Komarova and A. B. Malik, Annual Review of Physiology, 2010, 72, 463-493.
3. A. Lossinsky and R. Shivers, Histology Histopathology, 2004, 19, 535-564.
4. U. Bickel, Neurotherapeutics, 2005, 2, 15-26.
5. F.-R. E. Curry and T. Noll, Cardiovascular Research, 2010, 87, 195-197.
6. A. Chavez, M. Smith, D. Mehta and W. J. Kwang, in International Review of Cell and Molecular Biology, Academic Press, 2011, pp. 205-248.
7. C. C. Michel and F. E. Curry, Physiological Reviews, 1999, 79, 703-761.
8. R. F. Tuma, W. N. Durán and K. Ley, Microcirculation, Elsevier/Academic Press, 2008.
9. A. R. Pries, T. W. Secomb and P. Gaehtgens, Cardiovascular Research, 1996, 32, 654-667.
10. H. D. Intengan, G. Thibault, J. S. Li and E. L. Schiffrin, Circulation, 1999, 100, 2267-2275.
11. D. Mehta and A. B. Malik, Physiological Reviews, 2006, 86, 279-367.
12. J. R. Fox and H. Wayland, Microvascular Research, 1979, 18, 255-276.
13. R. K. Jain, J. M. Weissbrod, J. Wel, K. George and W. Sidney, in Advances in Cancer Research, Academic Press, 1980, pp. 251-310.
14. B. M. Fu, R. H. Adamson and F.-R. E. Curry, Journal of Biomechanical Engineering, 2005, 127, 270-278.
15. P. M. A. van Haaren, E. VanBavel, H. Vink and J. A. E. Spaan, Am J Physiol Heart Circ Physiol, 2003, 285, H2848-2856.
16. P. M. A. van Haaren, E. VanBavel, H. Vink and J. A. E. Spaan, American Journal of Physiology-Heart and Circulatory Physiology, 2005, 289, H2503-H2507.
17. J. A. Cooper, P. J. Del Vecchio, F. L. Minnear, K. E. Burhop, W. M. Selig, J. G. Garcia and A. B. Malik, Journal of Applied Physiology, 1987, 62, 1076-1083.
18. P. J. Gaillard and A. G. de Boer, European Journal of Pharmaceutical Sciences, 2000, 12, 95-102.
19. Y. Zhang, C. S. W. Li, Y. Ye, K. Johnson, J. Poe, S. Johnson, W. Bobrowski, R. Garrido and C. Madhu, Drug Metabolism and Disposition, 2006, 34, 1935-1943.
37
20. E. W. K. Young, M. W. L. Watson, S. Srigunapalan, A. R. Wheeler and C. A. Simmons, Analytical Chemistry, 2010, 82, 808-816.
21. R. Booth and H. Kim, Lab on a Chip, 2012.
22. P. A. Vogel, S. T. Halpin, R. S. Martin and D. M. Spence, Analytical Chemistry, 2011, 83, 4296-4301.
23. A. Gunther, S. Yasotharan, A. Vagaon, C. Lochovsky, S. Pinto, J. Yang, C. Lau, J. Voigtlaender-Bolz and S.-S. Bolz, Lab on a Chip, 2010, 10, 2341-2349.
24. C. Lochovsky, S. Yasotharan and A. Gunther, Lab on a Chip, 2012, 12, 595-601.
25. S. S. Bolz, L. Vogel, D. Sollinger, R. Derwand, C. Boer, S. M. Pitson, S. Spiegel and U. Pohl, Circulation, 2003, 108, 342-347.
26. G. Kaley, B. M. Altura and B. W. Zweifach, Microcirculation, University Park Press, 1980.
27. J. M. Tarbell, Cardiovascular Research, 2010.
28. R. Minshall, C. Tiruppathi, S. Vogel and A. Malik, Histochemistry and Cell Biology, 2002, 117, 105-112.
29. G. E. Simionescu N Fau - Palade and G. E. Palade, 1971.
30. L. J. Nugent and R. K. Jain, American Journal of Physiology - Heart and Circulatory Physiology, 1984, 15.
31. L. E. Gerlowski and R. K. Jain, Microvascular Research, 1986, 31, 288-305.
32. L. T. Baxter and R. K. Jain, Microvascular Research, 1988, 36, 108-115.
33. B. E. Steinberg, N. M. Goldenberg and W. L. Lee, Antiviral Research, 2012, 93, 2-15.
34. P. Claude, Journal of Membrane Biology, 1978, 39, 219-232.
35. M. N. Hart, L. F. VanDyk, S. A. Moore, M. Shasby and P. Cancilla, J. Neuropathol. & Exp. Neurol., 1987, 46, 141-153.
36. M. Inc., Billerica, MA, 2009.
37. W. P. I. Inc., Sarasota FL, May 2011.
38. M. Deli, C. Ábrahám, Y. Kataoka and M. Niwa, Cellular and Molecular Neurobiology, 2005, 25, 59-127.
39. F. Curry and R. Adamson, Annals of Biomedical Engineering, 2012, 40, 828-839.
40. D. Huh, B. D. Matthews, A. Mammoto, M. Montoya-Zavala, H. Yuan Hsin and D. E. Ingber, Science, 2010, 328, 1662-1668.
38
41. I. Barbulovic-Nad, S. H. Au and A. R. Wheeler, Lab on a Chip - Miniaturisation for Chemistry and Biology, 2010, 10, 1536-1542.
42. P. J. Lee, P. J. Hung and L. P. Lee, Biotechnology and Bioengineering, 2007, 97, 1340-1346.
43. K. S. Lee, P. Boccazzi, A. J. Sinskey and R. J. Ram, Lab on a Chip, 2011, 11, 1730-1739.
44. S. Chung, R. Sudo, V. Vickerman, I. Zervantonakis and R. Kamm, Annals of Biomedical Engineering, 2010, 38, 1164-1177.
45. D. Discher, C. Dong, J. J. Fredberg, F. Guilak, D. Ingber, P. Janmey, R. D. Kamm, G. W. Schmid-Schonbein and S. Weinbaum, Annals of Biomedical Engineering, 2009, 37, 847-859.
46. A. D. van der Meer, A. A. Poot, M. H. G. Duits, J. Feijen and I. Vermes, Journal of Biomedicine and Biotechnology, 2009, 2009.
47. S. S. Bolz, S. Pieperhoff, C. De Wit and U. Pohl, American Journal of Physiology-Heart and Circulatory Physiology, 2000, 279, H1434-H1439.
48. Y. N. Xia and G. M. Whitesides, Annual Review of Materials Science, 1998, 28, 153-184.
49. J. Hoefer, M. A. Azam, J. T. E. Kroetsch, H. Leong-Poi, M. A. Momen, J. Voigtlaender-Bolz, E. Q. Scherer, A. Meissner, S.-S. Bolz and M. Husain, Circulation Research, 2010, 107, 923-933.
50. S. Aldrich FITC-Dextran Data Sheet, 1997.
51. F. E. Curry, V. H. Huxley and R. H. Adamson, American Journal of Physiology - Heart and Circulatory Physiology, 1983, 245, H495-H505.
52. M. Burns-Bellhorn, R. Bellhorn and J. Benjamin, Invest. Ophthalmol. Vis. Sci., 1978, 17, 857-862.
53. M. W. R. Reed and F. N. Miller, Microvascular Research, 1988, 36, 104-107.
54. R. E. Rumbaut and A. J. Sial, Microcirculation, 1999, 6, 205-213.
55. F. N. Miller, D. E. Sims, D. A. Schuschke and D. L. Abney, Microvascular Research, 1992, 44, 166-184.
56. M. Steinbauer, A. G. Harris, C. Abels and K. Messmer, Langenbeck's Archives of Surgery, 2000, 385, 290-298.
57. R. E. Kalman, Journal of Basic Engineering, 1960, 82, 35-45.
58. J. Canny, IEEE Trans. Pattern Analysis and Machine Intelligence, 1986, 8, 679-698.
59. N. Moës, E. Béchet and M. Tourbier, International Journal for Numerical Methods in Engineering, 2006, 67, 1641-1669.
60. M. Mulder, Basic Principles of Membrane Technology Kluwer Academic Publishers, 1998.
39
61. L. K. Chin, J. Q. Yu, Y. Fu, T. Yu, A. Q. Liu and K. Q. Luo, Lab on a Chip, 2011, 11, 1856-1863.
62. A. Koller, The American journal of physiology, 1991, 260.
63. A. Koller and A. Huang, Hypertension, 1995, 25, 758-763.
65. D. Sun, A. Huang, E. H. Yan, Z. Wu, C. Yan, P. M. Kaminski, T. D. Oury, M. S. Wolin and G. Kaley, American Journal of Physiology - Heart and Circulatory Physiology, 2004, 286, H2249-H2256.
66. D. Montermini, C. P. Winlove and C. C. Michel, The Journal of Physiology, 2002, 543, 959-975.
67. W. N. Duran, J. W. Breslin and F. A. Sanchez, Cardiovascular Research, 2010, 87, 254-261.
68. E. A. Koop, M. F. B. G. Gebbink, T. E. Sweeney, M.-J. Mathy, H. F. G. Heijnen, J. A. E. Spaan, E. E. Voest, E. VanBavel and S. L. M. Peters, American Journal of Physiology - Heart and Circulatory Physiology, 2005, 288, H1218-H1223.
69. R. Furchgott, Circulation Research, 1983, 53, 557-573.
70. M. K. McKay, A. L. Gardner, D. Boyd and R. L. Hester, Hypertension, 1998, 31, 213-217.
71. J. B. Pawley, Handbook Of Biological Confocal Microscopy, Springer, 2006.
72. R. B. Bird, W. E. Stewart and E. N. Lightfoot, Transport Phenomena, J. Wiley, 2007.
73. W. Wang, M. J. Merrill and R. T. Borchardt, American Journal of Physiology - Cell Physiology, 1996, 271, C1973-C1980.
74. H. M. Wu, Q. Huang, Y. Yuan and H. J. Granger, American Journal of Physiology - Heart and Circulatory Physiology, 1996, 271, H2735-H2739.
75. D. Feng, J. A. Nagy, K. Pyne, I. Hammel, H. F. Dvorak and A. M. Dvorak, Microcirculation, 1999, 6, 23-44.
76. Y. S. Chang, L. L. Munn, M. V. Hillsley, R. O. Dull, J. Yuan, S. Lakshminarayanan, T. W. Gardner, R. K. Jain and J. M. Tarbell, Microvascular Research, 2000, 59, 265-277.
77. K. R. Lee, N. Kawai, S. Kim, O. Sagher and J. T. Hoff, Journal of Neurosurgery, 1997, 86, 272-278.
78. T. J. Williams and J. Morley, Nature, 1973, 246, 215-217.
79. S. Yasotharan, in Mechanical and Industrial Engineering, University of Toronto, Toronto, 2012.
80. E. B. Brown, Y. Boucher, S. Nasser and R. K. Jain, Microvascular Research, 2004, 67, 231-236.
40
81. K. Braeckmans, K. Remaut, R. E. Vandenbroucke, B. Lucas, S. C. De Smedt and J. Demeester, Biophysical Journal, 2007, 92, 2172-2183.
82. S. M. Weis and D. A. Cheresh, Nature, 2005, 437, 497-504.
83. J. Gavard and J. S. Gutkind, Nat Cell Biol, 2006, 8, 1223-1234.
84. W. L. Monsky, D. Fukumura, T. Gohongi, M. Ancukiewcz, H. A. Weich, V. P. Torchilin, F. Yuan and R. K. Jain, Cancer Research, 1999, 59, 4129-4135.
85. A.-K. Olsson, A. Dimberg, J. Kreuger and L. Claesson-Welsh, Nat Rev Mol Cell Biol, 2006, 7, 359-371.
86. D. O. Bates, Cardiovascular Research, 2010, 87, 262-271.
41
A.1 Microfluidic Device Design
The design of the microfluidic device was based on a previous iteration used in the first publication
of this technology by Günther et al..23 The design was carried out on Autocad (Autocad 2011, San
Rafael, CA, U.S.A), and was designed for:
i) An additional superfusion inlet for the use of ACh
ii) An additional perfusion inlet for introducing the fluorescent marker without
affecting the MOPS buffer flow
iii) Compatibility with the 10-inlet/outlet manifold provided by Quorum
Technologies (Quorum Technologies, Guelph, ON, Canada).
inlets, 1 superfusion outlet, 2 vacuum lines, 2 fixation lines and a loading well. Locations can be seen
in Fig. 2.1..
42
A.2 Post-Processing Script for Confocal Images
The following post-processing scripts were written for use with Matlab R2011a.
Script 1: Process confocal images and output intensities and concentrations
clear variables % enter number of frames imagenum=; % enter total time totaltime=329.593; %calibration um/pixels calibration=0.621; %Linear curve parameters obtained from calibration experiment conc_a=7.04229e3; conc_b=5.74697e-3; %y = 7.04229E+03x + 5.74697E-03 %Directories for outputting files mkdir('Output\Date'); directory='C:\Output\'; directory=strcat(directory,'Date\'); %read brightfield image BF = imread('Figure 3 bf_C002T002.tif'); %Figure and Plot Titles and Fileneames figure_title='Date - T1'; plot_filename='Date - T1'; fileto_cat='Date('; fileto_cat_type=').tif'; %Adjust contrast of brightfield image BF = imadjust(BF); %Perform Canny-Edge detection on Brightfield Image BW = edge(BF,'canny',[.1 0.4],6.6); %Show brightfield image and wait for user to define ROI imshow(BF); %ROI=[xmin ymin width height] ROI_1=getrect(); close all; if (ROI_1(1)<1) ROI_1(1)=1; end if (ROI_1(2)<1) ROI_1(2)=1; end if (ROI_1(1)+ROI_1(3)>512) ROI_1(3)=512-ROI_1(1); end %Define bounds of ROI to analyze xmin=round(ROI_1(1)); xmax=round(ROI_1(1))+floor(ROI_1(3)); ymin=round(ROI_1(2)); ymax=round(ROI_1(2))+floor(ROI_1(4)); %Create a mask to only take certain pixels Canny_image=BW(ymin:ymax,xmin:xmax); %determine the bounds of the region sze=size(Canny_image);
43
tempMASK=zeros(sze(1),sze(2)); %Determine location of vessel wall and set it as y_min in each row for column=1:sze(2) row=1; flag=0; while(row<=sze(1)) tempMASK(row,column)=0; if (flag==1) %set mask to be 1 below the wall tempMASK(row,column)=1; end if (Canny_image(row,column)==1) %store locations of wall flag=1; end row=row+1; end end row=row-1; MASK=zeros(512,512); MASK(ymin:ymax,xmin:xmax)=tempMASK;
%Store intensities of confocal images into array called SUM %SUM(x,y,frame number)=intensity %SUM_v2(x,y,frame number)=concentration SUM=zeros(512,512,imagenum); SUM_v2=zeros(512,512,imagenum); for frames=1:imagenum imagename=strcat(fileto_cat,num2str(frames),fileto_cat_type); temp=imread(imagename); temp=im2double(temp); %Convert intensities to concentrations temp_2=(temp-conc_b)/(conc_a); intensity(frames)=0.0; intensity_v2(frames)=0.0; %intensity is a matrix for intensities %intensity_v2 is a matrix for concentrations count=0; for r=1:512 for c=1:512 if MASK(r,c)==1 count=count+1; SUM(r,c,frames)=temp(r,c); SUM_v2(r,c,frames)=temp_2(r,c); intensity(frames)=intensity(frames)+temp(r,c); intensity_v2(frames)=intensity(frames)+temp_2(r,c); end end end %intensity_avg is a matrix for storing intensities/(pixels measured) intensity_avg(frames)=intensity(frames)/count; intensity_v2_avg(frames)=intensity_v2(frames)/count; end
%Take average intensity for frames and standard deviation avg_intensity=mean(intensity); std_intensity=std(intensity);
44
%Remove frames if outside standard deviation for i=1:imagenum a=abs(avg_intensity-intensity(i)); if (a<std_intensity) take_frame(i)=1; else take_frame(i)=0; end end %Output frame numbers that are to be included for measurement take_frame framecount=0; for i=1:imagenum if (take_frame(i)==1) framecount=framecount+1; end end %Take total intensities of pixels within ROI and outside of vessel wall %for all frames to be included tmp_1_values=zeros(sze(1),sze(2)); tmp_2_values=zeros(sze(1),sze(2));
for r=ymin:ymax count=0; tmp_1=0; tmp_2=0; rowcount(r-ymin+1,1)=0; for frame=1:imagenum for c=xmin:xmax if (take_frame(frame)==1) tmp_1_values(r-ymin+1,c-xmin+1)=SUM(r,c,frame)+tmp_1; tmp_1=SUM(r,c,frame)+tmp_1; tmp_2_values(r-ymin+1,c-xmin+1)=SUM_v2(r,c,frame)+tmp_2; tmp_2=SUM_v2(r,c,frame)+tmp_2; count=count+1; rowcount(r-ymin+1,1)=rowcount(r-ymin+1,1)+1; end end end end %Calculate average intensities for each column for r=ymin:ymax row_avg_1(r-ymin+1,1)=mean((tmp_1_values(r-ymin+1,:)/(rowcount(r-
ymin+1,1))),1,2); end max_value1=max(row_avg_1); max_value2=max(row_avg_2);
%Store values into a normalized intensity matrix row_n row_n=row_avg_1/max_value1; a=1;
45
for i=1:size(row_avg_1,1) dist1(a,1)=calibration*a; a=a+1; end
%Mark the last point where the intensity is =1 (x_1) x_1=1; for i=1:row if row_n(i,1)==1 x_1=i; end end %Create arrays for outputting intensities and concentrations row_plot_1=row_avg_1(x_1:row,1); row_plot_std_1=row_std_1(x_1:row,1); row_plot_2=row_avg_2(x_1:row,1); row_plot_std_2=row_std_2(x_1:row,1);
a=1; %Create distance vector for i=1:size(row_plot_1,1) dist2(a,1)=calibration*a; a=a+1; end
%Output values into TXT files dlmwrite(strcat(directory,plot_filename,'-raw intensities.txt'), [dist1
row_plot_2 row_plot_std_2], 'delimiter', '\t'); %Plot concentration profiles and save plot as a PNG h=figure; errorbar(dist2,row_plot_2,row_plot_std_2,'or'); xlabel('Distance from vessel wall (um)'); ylabel('Concentration (M)'); title(strcat('Concentration Profiles - ',figure_title)); %legend(legend1,legend2); print(h,'-depsc','-tiff','-r300',strcat(directory,plot_filename,'.eps')); print(h,'-dpng','-r300',strcat(directory,plot_filename,'.png')) close(h);
46
Script 2: Take concentration profiles and carry out a linear fit
%Prior to starting, concentration profiles show be saved as matrix A and
%organized with column headers:(distance,concentration value)
%Store values into matrices newmatrix_X for distance & %newmatrix_Y for concentration newmatrix_X=A(:,1)/1000; newmatrix_Y=A(:,2); loopcondition=size(newmatrix_X,1); %Set settings for minimum linear region %length span=8; %starting point start_position=5; start_frame=start_position+span; %Perform linear fits for all the points starting from starting point and %store the square of the residuals for i=start_frame:loopcondition %if condition for minimal residual
i),'poly1'); res(i-start_frame+1,1)=gof.rsquare; end %Find the best fit according to highest r-squared value [y,x]=max(res(:,1)); value(1)=y; location(1)=x+start_frame-1; %Perform a fit for the span with the highest r-squared value, and compute %the values for the fit function
tion:location(1)),'poly1'); Y=feval(fobj,newmatrix_X(start_position:location(1))); %Output the slope data for use with the results of the numerical model slope_data(1,1)=fobj.p1 slope_data(1,2)=fobj.p2
Results from script 2 are used with results from the numerical model correlation curve (Fig. 3.4b) to
obtain permeability coefficients.
47
A.3 Numerical Model
The numerical model developed and outlined in Section 2.5 was used as a solver of the 2-D
convective and diffusive transport equations for our experimental problem. Concentration
distributions as a function of position and time were solved for and used to obtain experimental
permeability coefficients. In developing the model, mesh dependence was determined as a balance
between computational costs and solution accuracy (Fig. A.2).
Figure A.2: Mesh dependence of numerical model. a) Rate of change on concentration obtained at
vessel centerline for different mesh sizes. Mesh sizing was varied between Extra Fine and Extremely
Fine, with numbers of refinements in the lumen varied between low (1) and high (5). b) The resulting
mesh used throughout permeability experiments (extremely fine element size, with refinements in
the vessel wall).
A mesh size setting of Extremely Fine with low refinements in the area representing the vascular wall
resulted in 65500 elements, half the number of elements and an order of magnitude less of degrees
of freedom to solve for than an Extra Fine element size with high refinements. The mesh choice
resulted in a 4.05% difference in average computed values.
48
The developed model also served the purpose of determining experimental parameters to be used
during permeability experiments. Inlet flow, Qin, plays a vital role in determining the concentration
profile. Too high of a Qin would result in a thin boundary thickness, while too low of a Qin would
result in a long transient time to reach steady-state conditions. A parametric sweep for different
abluminal Peclet numbers, Pe, was carried out to understand this effect given our experimental
parameters (section 2.3).
Figure A.3: Effect of Pe number on the transient and steady-state solutions. Transient study of time
to steady-state conditions for flow from no flow lumen to superfusion channel for different Pe in
superfusion channel inlet: a) 10, 100, 1000, 10000 , b) 0.01, 0.1, 1. c) Effect of Pe on concentration
profiles along centre-line of vessel, measured from vessel wall in the axial direction.
49
A.4 VEGF as a Positive Control
The role of VEGF in affecting vascular permeability has been broadly researched using both in vitro
73, 76, 82, 83 and in vivo 74, 84 methods, so much so that initially VEGF-A was first discovered as vascular
permeability factor (VPF). VEGF is seen to act through a variety of mechanisms to induce
hypermeability, specifically, as described in Olsson et al.’s 2006 review:85 the forming of fenestrations
in the endothelium, the assembly of calveolae into vesiculovacuolar organelles and the induction of
trans-endothelial pores. VEGF regulation of permeability is dependent on the production of
eNOS.85, 86
VEGF was chosen as a positive control to further validate the presented experimental method.
Experiments were carried out to examine the effect of VEGF on EC and SMC functioning (Fig.
A.4a-b), followed by permeability experiments (Fig. A.5).
Figure A.4: VEGF effect on smooth muscle cell and endothelial cell viability. Dose response
measurements to a) PE and b) ACh for four sequential conditions, Static: after heating and
pressurization, no perfusion flow. Flow : 10 mins. of 0.3 μl/h perfusion flow (1.06 dynes/cm2) of
MOPS buffer. VEGF: 10 mins. of perfusion of MOPS + 1 nm VEGF (#V4512. Sigma-Aldrich,
Oakville, ON, Canada). Post-VEGF Flow: 10 mins of 0.3 μl/h perfusion flow of MOPS buffer.
50
Figure A.5: VEGF effect on permeability coefficient. Dose response measurements to a) PE and b)
ACh prior to and in between three permeability coefficient measurements c) Run 1: 0.3 μl/h
perfusion flow (1.06 dynes/cm2) of MOPS buffer + 9 mg/ml 4kDa FITC-Dextran, Run 2 (control):