1 Ultrashort laser bioengineering: Micropatterning of collagen fibers __________________ Alice Rebière 20160620 __________________ Master of Science Thesis in Engineering Physics at KTH Supervisor: Raphaël Devillard (INSERM/ALPhANOV) Cosupervisor: Dora AïtBelkacem (ALPhANOV) Examiner: Marina Zelnina KTH TRITA FYS 2016:25 ISSN 0280316X ISRN KTH/FYS/16:25—SE
Welcome message from author
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
1
Ultrashort laser bioengineering: Micropatterning of collagen fibers
__________________ Alice Rebière
2016-‐06-‐20 __________________
Master of Science Thesis in Engineering Physics at KTH
Supervisor: Raphaël Devillard (INSERM/ALPhANOV) Co-‐supervisor: Dora Aït-‐Belkacem (ALPhANOV)
Examiner: Marina Zelnina KTH
TRITA FYS 2016:25 ISSN 0280-‐316X ISRN KTH/FYS/-‐-‐16:25—SE
2
3
Acknowledgements: Je souhaite tout d’abord remercier Raphaël pour sa présence, sa gentillesse, son aide, son encadrement, le prêt de sa voiture, … Bref, merci d’avoir été un super encadrant dynamique, tout au long du stage, et de m’avoir supportée! Merci d’avoir cru en moi, et de m’avoir proposé de retravailler ensemble. Merci à Dora également pour l’encadrement, pour l’aide au quotidien à ALPhANOV, pour les bons moments en manip, l’aide sur le rapport, le thé au Yamato … Merci à Jérôme pour m’avoir initié à l’art de la fabrication des fibres, à la culture cellulaire, et merci pour le bizutage des débutants chimistes! Merci à Clémentine et Baptiste de m’avoir soutenu en fin de stage, et pour leur regard attentif sur la rédaction de mon rapport de stage. Merci à Simon pour l’entraide sur les manips des fibres, de la conception à l’usinage en passant par la caractérisation. Merci à Bruno pour son avis éclairé sur les simulations optiques, ainsi que le soutien pointu en optique. Merci à Murielle pour son soutien sur la partie biologique et pour tout l’apprentissage en culture cellulaire! Merci à ceux qui m’ont permis de passer des journées agréables dans les trois bureaux où j’ai eu l’occasion de travailler: d’abord la Meso team à ALPhANOV (Charly, Christophe, Guillaume et Anthony), puis la Nano team (Baptiste, Bastien – merci aussi pour l’aide sur les manip Tangerine!, Clémentine, Dora, Laura), et bien sûr à l’INSERM, Olivia (et sa puce!), Yoyo, Simon, Pauline et Camille. Merci également à l’ensemble de la BU MUL à ALPhANOV, et plus globalement, à l’ensemble du personnel d’ALPhANOV et de l’ unité BioTis de l’INSERM pour leur support tout au long de ces semaines passées si vite, et les bons moments passés ensemble.
4
Abstract The lack of organ donors for transplants is becoming a huge burden. In attempt to
palliate this shortage and to help tissues to self-‐repair, tissue engineering labs are developing bioengineered human tissues. One of the most challenging issues encountered in the field of tissue engineering is the vascularization of the bioengineered tissues. Several approaches have been developed to face this issue.
Among them, the approach developed in this project deals with the biofabrication of collagen-‐based microfibers to mimic blood capillaries that would be integrated inside a tissue-‐engineered construct. The realization of these capillaries used laser machining techniques applied to tissue engineering: endothelial cell-‐laden microfibers of collagen are produced with a diameter around 100 to 150 µm. The collagen core is then machined by ultrashort laser pulses using different wavelengths (532 and 1064 nm) and different pulse durations (picosecond and femtosecond).
In this study, we demonstrate the feasibility of the process. The precise machining process allowed the creation of 10 to 20 µm voids in the fiber-‐shaped construct without extended damage for surrounding cells. Confocal microscopy examination of the fibers demonstrated variation of the diameter of the hole regarding the machining setups and laser energies used.
These machined microfibers would be perfused to validate the mechanical characteristics and flow resistance of the tissue-‐engineered construct created. In a long-‐term, it is expected that these perfused microfibers can be used as an easy prevascularization method of tissue-‐engineered constructs and as an essential component of artificial organs.
5
Table of contents: Acknowledgements: ........................................................................................................................... 3
Abstract ................................................................................................................................................ 4
List of Figures: ...................................................................................................................................... 7
List of Tables: ....................................................................................................................................... 9
I. Context: tissue engineering ..................................................................................................... 10 A. Tissue Engineering ............................................................................................................................. 10 B. Vascularization of tissue substitutes ............................................................................................... 12 1. Vascularization: the most challenging issue of tissue engineering ......................................... 12 2. State of the art of vascularization solutions ...................................................................................... 13
C. Laser technology applied to tissue engineering ............................................................................ 16 1. Laser-‐matter interaction applied to biological tissues ................................................................. 17 2. Laser micropatterning for tissue engineering ................................................................................. 19
D. The project’s objectives .................................................................................................................... 20
II. Material and methods ............................................................................................................. 21 A. Preparation of the fibers ...................................................................................................................... 21 1. Cell culture ....................................................................................................................................................... 21 2. Collagen ............................................................................................................................................................. 21 3. Fiber ................................................................................................................................................................... 21 4. Fixation of fibers ........................................................................................................................................... 23
B. Micromachining the fibers ................................................................................................................... 23 1. The laser sources .......................................................................................................................................... 23 2. The optical set ups ....................................................................................................................................... 24 3. Sample-‐holding systems ............................................................................................................................ 26
C. Characterization method: confocal fluorescent microscopy .......................................................... 27 1. DAPI Labeling ................................................................................................................................................. 27 2. Live/Dead® Assay ....................................................................................................................................... 27
III. Results ....................................................................................................................................... 28 A. Influence of the laser parameters on the machining ................................................................... 28 1. Influence of the energy: threshold determination ......................................................................... 28 2. Influence of the pulse duration on the machining .......................................................................... 31 3. Influence of the wavelength on the machining ................................................................................ 31
B. Fabrication of capillary substitutes ................................................................................................. 32 1. Whole fiber machining ............................................................................................................................... 32 2. Cell viability ..................................................................................................................................................... 32
IV. Discussion ................................................................................................................................. 33 A. Influence of the laser parameters on the machining ................................................................... 33 1. Influence of the energy on the machining: threshold determination .................................... 33 2. Influence of the pulse duration ............................................................................................................... 34 3. Influence of the wavelength ..................................................................................................................... 35
B. Fabrication of blood capillaries substitutes ................................................................................... 35 1. Machining on the collagen ........................................................................................................................ 36 2. Size of the channels ...................................................................................................................................... 36 3. Whole fiber machining: choice of optimal settings ........................................................................ 36 4. Cell viability ..................................................................................................................................................... 37
6
Conclusion and Perspectives .......................................................................................................... 37
Annex 1: Example of a G-‐code program ................................................................................ 39
Annex 2: Tests with Sirius laser (green) on Eclipse 1: optimization of patterning set-‐up .... 40 A. From the sample holding system V2 to V3 .................................................................................... 40 B. Highlighting the need of Eclipse 2 set-‐up ....................................................................................... 42
References ....................................................................................................................................... 43
7
List of Figures: Figure 1: General principle of tissue engineering (3) ............................................................. 11 Figure 2: Schematic description of diffusion and transport processes in vascularized tissues in vivo [inspired from(9)] ............................................................................................................ 12 Figure 3: Critical distance between cells and blood vessels (11) ............................................ 13 Figure 4: Outline of the advantages and drawbacks of the different strategies for vascularization approaches (9) ............................................................................................... 14 Figure 5: Overview of the advantages and drawbacks of the different strategies for vascularization in tissue engineering (15)……………………………………………………………………………13 Figure 6: Scheme of nonlinear interaction phenomena with its timescale (courtesy to Dr John Lopez) ………………………………………………………………………………………………………………………………..16 Figure 7: Creation of plasma and optical breakdown in water ............................................... 18 Figure 8: Nonlinear absorptionfor ultrashort laser machining (courtesy of Dr john Lopez) ... 19 Figure 9: A) Creation of the sheath around the glass capillary B) Replacement of the glass capillary by the collagen/cell solution (38) ............................................................................. 22 Figure 10: Fiber with (on the left) and without (on the right) the sheath. The green part is the collagen, on the edge there are the cells (38) ........................................................................ 23 Figure 11: Eclipse I inverted microscope, Aerotech stages and sample ................................. 24 Figure 12: Scheme of the Eclispse 1 set-‐up ............................................................................. 25 Figure 13: Scheme of the Eclipse II set-‐up .............................................................................. 25 Figure 14: Photo of the Eclipse II set-‐up ................................................................................. 26 Figure 15: Confocal image of a fiber machined at 2.2 µJ: three views (a) DAPI labelling in the cells (b) autofluorescence of the collagen (c) merge of the two images (cells and collagen) 29 Figure 16: Confocal image of the fiber machined at 3 µJ (merge of the two images) ............ 29 Figure 17: Zoom on a part of the fiber (a) DAPI label on D1 cells (b) autofluorescence of the collagen (c) merge of the two previous images ...................................................................... 30 Figure 18: Zoom on a part of the fiber (a) DAPI label on D1 cells (b) autofluorescence of the collagen (c) merge of the two previous images ...................................................................... 30 Figure 19: Zoom on a part of the fiber (a) DAPI label on D1 cells (b) autofluorescence of the collagen (c) merge of the two previous images ...................................................................... 31 Figure 20: Whole fiber machined by Satsuma laser (IR) ......................................................... 32 Figure 21: Live/dead protocol on a non-‐fixed fiber of endothelial cells machined at 3 µJ: observation at the confocal microscope (green: live cells, red: dead cells). This live/dead assay is representative for n=6 fibers ..................................................................................... 33 Figure 22: Live/dead assay on a non-‐fixed fiber of endothelial cells machined at 3 µJ: observation on the confocal microscope (green: live cells, red: dead cells), on thee different sections (longitudinal and transverse sections) ...................................................................... 33 Figure 23: Absorption coefficients of several species including collagen (collagen fibrils) and water (39) ............................................................................................................................... 35 Figure 24: Microscope observation of the fiber machined at 4,48 µJ (obj 10x) (a) before (b) after the machining occured ................................................................................................... 40 Figure 25: Confocal observation of the fiber machined at 4,48 µJ (three different sections) 41 Figure 26: Different fibers machined at 2,24 µJ (a) Fiber machined (b) Exploded fiber (c) Torn fiber (d) Void bubbles appearance in the fiber ....................................................................... 41 Figure 27: Point Spread Function of the beam with an agarose gen layer of 400 µm (a) and without (b) the agarose gel layer ............................................................................................ 42
8
9
List of Tables:
Table 1: Characteristics of the different tests performed .......................................................... 26 Table 2: Parameters used in Sirius (IR) tests ............................................................................................. 28 Table 3: Energy tests on Sirius (IR) ............................................................................................................... 28 Table 4: Size of the channels depending on the energy for a machining with Sirius laser
(picosecond). .............................................................................................................................................. 29 Table 5: Parameters used in Satsuma (IR) tests ...................................................................................... 29 Table 6: Energy tests on Satsuma (IR) ......................................................................................................... 30 Table 7: Size of the channels depending on the pulse duration for a machining at 3 µJ .......... 31 Table 8: Characteristics of the lasers used for wavelength-‐dependent test ................................. 31 Table 9: Optimal parameters ......................................................................................................................... 37 Table 10: Settings of Sirius (green) laser for tests on wavelength effect ........................................ 40 Table 11: Energy tests on Sirius (green) laser ........................................................................................... 40
10
I. Context: tissue engineering A growing number of people are suffering from bone diseases such as osteoporosis,
pseudo-‐arthritis or osteosarcomas, consisting of a major health issue. In industrialized countries, the population aging is the main factor to explain that growth. The Health High Authority (Haute Autorité de Santé) in France declares that 42778 bone substitutes have been used in France in 2011 for 10 million euros. More than two million bone grafts are performed annually around the globe in order to remedy to the bone defects in orthopedics and dentistry surgery.
In clinical application, the therapeutic solutions are mainly centered on the use of human, animal or synthetic transplants. For bone tissue, the use of auto transplantation remains the ‘gold standard’ as it represents the ‘ideal’ material but it is limited in quantities and can be the cause of infections, pains, and morbidity on the donor site. The main risks relative to allotransplantation and xenotransplantation are the rejection of the transplant and the transmission of pathogenesis agents. As for the synthetic materials used in clinical applications, their lack of biological properties remains a major disadvantage to their use and illustrates the limitations of such substitutes for tissue regeneration. It is especially difficult to recreate tissue architecture, as well on the macroscopic level (general shape), as on the microscopic level (3D microorganization of cells and other tissue components).
The possibility to generate new tissue substitutes with optimized properties is therefore
a major stake and a real clinical need (1). This is what the field of tissue engineering tends to achieve.
A. Tissue Engineering
Tissue engineering has been defined as “the development of biological substitutes that enable to restore, maintain, or improve tissue and organ functions to avoid using a mechanical system, organ transplantation or surgical reconstruction” (2). Van Blitterswijk precises that tissue engineering gathers “all the techniques and methods inspiring from engineering and life sciences, used to develop biological substitutes”(3). That research field allows to consider new therapeutic solutions where you expand, manipulate and handle cells in order to seed them into a construct that can further be implanted (Fig. 1).
11
Figure 1: General principle of tissue engineering (3)
Within the tissue engineering field, several categories can be distinguished. Biofabrication is one of them: it is defined as “the automated generation of biologically functional products with structural organization from living cells, bioactive molecules, biomaterials, cell aggregates such as microtissues, or hybrid cell-‐material constructs, through bioprinting or bioassembly and subsequent tissue maturation processes” (4) Biofabrication itself gathers two techniques: bioassembly and bioprinting.
Bioassembly can be defined as “the fabrication of hierarchical constructs with a
prescribed 2D or 3D organization through automated assembly of pre-‐formed cell-‐containing fabrication units generated via cell-‐driven self-‐organization or through preparation of hybrid cell-‐material building blocks, typically by applying enabling technologies, including microfabricated molds or microfluidics” (4). This concerns all the tissue-‐engineered constructs where scaffolds and cells are assembled together.
Bioprinting is “the use of computer-‐aided transfer processes for patterning and
assembling living and non-‐living materials with a prescribed 2D or 3D organization in order to produce bio-‐engineered structures serving in regenerative medicine, pharmacokinetic and basic cell biology studies” (4). There are several kind of bioprinters such as the ink-‐jet bioprinter, the microextrusion bioprinter or the laser-‐assisted bioprinter.
3D bioprinting allows printing and organizing cells likely as in a cellular tissue. Given
the actual state of the art in this field, 3D bioprinting allows printing cells recreating parts of organs which could mainly be used by the industry in order to develop personalised and adapted treatments to patients (5).
One of the tissue engineering applications this project is focused on is bone tissue
regeneration, which especially uses laser assisted bioprinting. The losses of bone tissues after a trauma represent a huge challenge. Clinically, some solutions have been developed to help bone tissue repair such as grafts (autotransplants, allotransplants and xenotransplants) and metallic bone substitutes (ceramic, polymers and composites). The therapeutic strategies now turn towards bone tissue engineering which raises an economic and scientific interest mainly in orthopedics surgery, reconstruction surgery and oromaxillofacial surgery.
12
As in grafting or wound healing, vascularization is one of the key limitations of the use of engineered bone tissue.
B. Vascularization of tissue substitutes
1. Vascularization: the most challenging issue of tissue engineering In regenerative medicine applications, three main challenges have to be overcome. The first one is the creation of biomaterials able to be transplanted into a body without
being rejected. Biocompatible materials are now handled much more easily: they can be braided, knitted, all kinds of scaffold constructs can be created, especially with the growth of 3D printing technologies (6). The main requirements for a scaffold in tissue engineering are easy cell penetration in the scaffold, a good cell distribution and cell proliferation. Besides, the scaffold should have a good permeability to culture medium, should be easily in vivo vascularized once implanted. The maintenance of cell phenotypes, the adequate mechanical properties are as important as a controlled biodegradation and an easy fabrication of the scaffold (6) .
The second main challenge is the difficulty to grow enough cells in vitro, out of the original body. During the last ten years scientists managed to develop many different cell types, especially stem cells like induced pluripotent stem cells (7). But there are still some types of cells that cannot be grown directly from a patient (liver cells, pancreas cells, nervous cells…) (8).
Finally, the third issue is the vascularization issue which is the ability to provide sufficient blood supply to the tissues and organs in order to enable them to survive during the initial phase after implantation (9)
The vascularization of tissues allows the cells within a tissue to get the correct amount of oxygen and nutrients they need to live. It also allows them to get rid of the waste and carbon dioxide they produce (Fig. 2).
Figure 2: Schematic description of diffusion and transport processes in vascularized tissues in vivo [inspired
from(9)]
Nevertheless, for a tissue or for a tissue-‐engineered construct to grow beyond 100-‐200 µm (the maximum distance the oxygen can browse within cells), new blood-‐vessel formation
13
is required [(9), (10)]. Otherwise, the cells situated further than 200 µm from a blood capillary will die (Fig. 3).
Figure 3: Critical distance between cells and blood vessels (11)
During in vitro culture, larger tissue-‐engineered constructs can be supplied with nutrients in perfusion bioreactors for example (12). After the implantation of the tissue construct, the supply of nutrients is limited by the average range of the diffusion processes.
The lack of vascularization in current tissue constructs is the main limit to the use of biomaterials and tissue substitutes during the implantation process (13). The only viable tissue engineered clinical products today are limited to very thin and poorly vascularized tissues such as the epidermis part of the skin or cartilage (14). Bigger grafts are not vascularized enough leading to insufficient nutrients and oxygen exchanges, which finally turns into cell necrosis and rejection of the graft (14,15).
To allow bigger tissue constructs to live, a homogeneous vascularization is needed: a
capillary network should be formed to deliver the proper amount of nutrients needed by the tissue-‐engineered construct. Once implanted, the implanted cells secrete signals as a reaction to hypoxia. These signals are received by the implanted host’s blood vessels which will invade the implant (15). However, The spontaneous growth of vessels is limited to limited to several tenths of micrometres per day so the complete vascularization of an implant would take a few weeks (16). In the meantime, an insufficient vascularization would lead to hypoxia deeper in the tissue and nutrient deficiencies.
2. State of the art of vascularization solutions
Several strategies are implemented to enhance vascularization. They can be sorted in cell-‐based techniques (based on the capacity of endothelial cells to form new vessels) and scaffold-‐based techniques (focused on the generation of vessel structures via scaffold creation). Both present advantages and drawbacks (Fig. 4 and Fig. 5).
14
Figure 4: Outline of the advantages and drawbacks of the different strategies for vascularization approaches
(9)
Figure 5: Overview of the advantages and drawbacks of the different strategies for vascularization in tissue
engineering (15)
a) Cell-‐based strategies Cell-‐based strategies rely on two processes of vessel formation: angiogenesis
(sprouting of capillaries from pre-‐existing blood vessels) and in vivo vasculogenesis (the assembly of undifferentiated endothelial cells to capillaries). Vasculogenesis is still possible
15
after birth. Endothelial progenitor cells differentiate to endothelial cells, which grow and organize to create first primitive vessel networks (17).
The prevascularization of tissue constructs with networks of capillary aims at reducing the time spent to connect the host tissue and the implantation (9).
For in vivo prevascularization, a non-‐vascularized tissue-‐engineered construct is
implanted into a region with a blood vessel suitable for microsurgical transfers. This way, a transplantable macrovessel is close enough to the graft. Within several weeks, the vascularization will grow inside the tissue-‐engineered construct and will form a microvascular network. After this step, the tissue-‐engineered construct is re-‐implanted at the defect site (18). However, this strategy needs three different surgeries (one to implant the construct at the vascularization site, one for the removal and another one to implant the construct at the final defect site) which may consist in a huge drawback.
Cell sheet engineering is another solution that deals with living cells growing in
biomaterials: confluent monolayers of cells are transplanted to ischemic tissues and can be layered to create a 3D construct: successful revascularization has been achieved this way (19–21).
The capacity of the endothelial cells to form prevascular structures in specific culture conditions is used for in vitro prevascularization. The in-‐vitro prevascularization strategy combines the endothelial cells with different types of cells (myoblasts or fibroblasts) by placing them in co culture in order to result in the formation of a prevascular network within the tissue (9).
The objective is to build a 3D prevascularized structure. After the implantation, the prevascularization network anastomoses (connects) to the vascular network of the host: the host blood vessels will only grow in the external regions of the construct until they meet the prevascular network. This way, the time needed for a complete vascularization is greatly reduced from weeks to days (15,22). Nevertheless, the perfusion is not as fast as with in vivo prevascularization because the vascular network is not microsurgically connected after the implantation.
The lack of microsurgical connections is a huge issue for in vitro prevascularization so the goal of future works would be to include a vascular network creation in the in vitro tissue-‐construct to make the anastomose faster.
Another strategy to induce neovascularization in vitro is angiogenic factor delivery.
Indeed the angiogenic factors help to improve the vascularization after the implantation (23). The angiogenic factors are delivered to stimulate the formation of blood vessels (Vascular Endothelial Growth Factor (VEGF) and bFibroblast Growth Factors (bFGF) for example), to stabilize the new vessels formed, but also for indirect approaches with the delivery of other factors that stimulate the cells situated near the vascularization site to produce angiogenic factors. However, the use of angiogenic factors (for example VEGF) might negatively influence the differentiation of the neighbouring cells that are present on co culture sites (for example, endothelial differentiation instead of osteoblast differenciation in the outskirt of the site);
16
therefore they should be used as little as possible. Besides, the use of such factors is not really satisfying yet as the resulting vessels are quite disorganized, or even leaky (5).
b) Scaffold-‐based strategies The scaffold–based strategy requires a good architecture and design of a scaffold: for example a scaffold with bigger pores will enable faster vessel ingrowth (15). The pores also have to be interconnected to allow cell migration and thus vascularization. The 3D scaffolds can be prepared via layer-‐by-‐layer assembly, from a wide range of choice of materials. Two types of scaffolds are mainly used: the biologically-‐derived vascularized scaffolds and the synthetically manufactured vascularized scaffolds.
The reuse of biological structures is very promising in the biologically derived vascularized scaffolds. The decellularization of mammalian sections allows to provide a 3D structure supplying microvascular networks. Several tests have been conducted on small intestinal submucosa of pigs (9). A whole-‐heart scaffold with intact geometry and vasculature has been constructed by Taylor et al (24). The advantages provided by these naturally-‐derived scaffolds are the biocompatible matrix, and the provision of relevant geometries (25). However, the lack of standardization and reproducibility stands as a limitation of this method, as well as the difficulty to get rid of all the cells that could induce an immune reaction and the need to recolonize the scaffold with another type of cells. On the contrary, the use of synthetically manufactured vascularized scaffolds allows a high reproducibility in the tests realized. The choice of the material and all the settings including scaffold stiffness, surface topography, structural qualities, biocompatibility, porosity offers a wide range of opportunities. Some of the most-‐used are naturally-‐derived polymers such as collagen, gelatine, or hyaluronan give some biocompatibility to the material. The degradability of the material is also a choice criterion as healing and degradation of the biomaterial need to be concomitant.
c) Towards a global approach
The final goal of a successful approach would be not to focus on one of these strategies, but investigate the integration of several in order to combine their assets and eliminate their weaknesses. In the end, the important thing is the number of functional vessels and the amount of blood cells they can carry.
In order to overcome the limitation that the lack of vascularization represents and to
develop larger tissue substitutes that can be implanted, research is focusing on the development of in vitro vascular structures for new regenerative therapies (27). Besides, the lack of implanted-‐material vascularization can be overtaken by the reproduction of the local microenvironment of cells. This allows to mime tissue complexity and promote cell interactions (28).
C. Laser technology applied to tissue engineering
Since the first Laser was created in 1960, laser technology has kept finding more and more new applications. As the French researcher Pierre Aigrain said, “we are used to having
17
a problem and looking for a solution. In the case of the laser, we already have a solution and we are looking for the problem”. Lasers are now commonly used in the field of health and biomedical engineering for diagnostic and therapeutic issues: they are a faster less-‐invasive more-‐precise method than other traditional and surgical methods. A broad range of applications have appeared including dermatology, surgery, odontology, ophthalmology, gastroenterology, urology, gynecology, cardiology, neurology, cell biology, etc.
Lasers are also spreading in the field of tissue engineering. The first example is the Laser-‐Assisted Bioprinting (LAB) technology (29,30). Another example is the laser use in the field of bioassembly for micropatterning applications (31–34).
1. Laser-‐matter interaction applied to biological tissues
A huge variety of applications in biology and in the medical field has appeared with the growth of laser technology. This range of applications is available thanks to the many parameters depending on lasers: they have different wavelengths, pulse duration, pulse energy, powers, or even focalization conditions, which allow to use them on different kind of tissues (for instance the eye, the skin, the surface of an organ, more indepth focalization…). The wavelengths allow different kind of interactions: the ultraviolet lasers can interact directly with molecular bonds ; the far infrared wavelengths can generate thermal effects, whereas in the visible and near infrared laser can be optimized to maximize the depth of penetration for example in the ocular tissues. The pulse duration influences the laser-‐tissue interaction (See Fig 6).
In this project, the goal was to perform some intra-‐volume machining into fibers made of a transparent collagen hydrogel. For this, the processes used are nonlinear interaction regimes such as photo disruption. Therefore, ultrashort lasers have been used (with pulse durations in the picosecond range and under). With pulse duration shorter than the thermal diffusion characteristic time, an almost athermal and very localized effect can be achieved in the tissue: an optical breakdown created which is a photodisruption effect. This way, the surrounding tissues are not damaged.
When the laser beam is focalized enough and with a high enough average power, multiphotonic absorption (MPA) is observed (Fig 6).
Figure 6: Scheme of nonlinear interaction phenomena with its timescale (courtesy to Dr John Lopez)
18
The energy gap of tissue is much higher than the energy brought by a single photon which is not enough for valence electrons to go to the conduction band. Therefore several photons are needed to excite the electrons enough to make them change bands.
Multiphotonic absorption is a nonlinear phenomenon possible with high intensity. This phenomenon is even more important when the pulse duration decreases: a higher number of photons will be concentrated in a shorter time and will therefore increase the intensity and the probability of nonlinear absorption (36). With a pulse duration in subpicosecond order, a few microjoules per pulse should be enough to create a local modification in a transparent material.
Once the electron is excited, it keeps interacting with the incident laser beam and its energy will increase (inverse Bremsstrahlung phenomenon). Its agitation increases with this energy and it finally collides with another electron. The collision will transfer some of the energy to the second enabling it to be transferred to the conduction band (secondary electron), and so on. The first electron with a high enough energy will remain in the conduction band while the secondary electrons are heated by the electromagnetic radiation of the laser (by inverse Bremsstrahlung phenomenon). They will therefore collide with the valence electrons. This creates avalanche ionization. The density of the electron cloud will increase until the formation of a plasma. When the pressure of the plasma gets superior to the tension strengths that maintain the collagen fibrils together, the threshold of the optical breakdown is achieved. A cavitation bubble is formed, then resorbs creating disruption in the focalization zone of the tissue. In tight focusing conditions, the modification zone has a comet shape, where the photoexcited volume is embedded into the heat affected zone. An example of the modification zones that can occur is displayed in Figure 7: (a) τ = 6ns, λ = 1064nm, E = 8,2 mJ (b) pulse duration τ = 30ps , λ = 1064nm , E = 740 μJ (c) pulse duration τ = 100fs, λ = 580nm, E = 35 μJ (37) .
Figure 7: Creation of plasma and optical breakdown in water
The machining process is described on Figure 8.
19
Figure 8: Nonlinear absorptionfor ultrashort laser machining (courtesy of Dr john Lopez)
2. Laser micropatterning for tissue engineering
The scaffolds used in tissue engineering techniques are often made of hydrogels. Hydrogels mainly consist of water, and can be used as guiding structures for cellular growth, differentiation, and regeneration. Patterning hydrogels for further cell seeding allows to keep the cells hydrated enough during the process thanks to the water contained in the hydrogel. Besides, as hydrogels are permeable to nutrients, the cells receive all the oxygen and nutrients they need to live and prolifer.
Several works on hydrogel micropatterning have been reported in the literature. All can be useful for vascularization issues in tissue engineering.
One of the most used hydrogel is collagen. Collagen is the most abundant protein in mammals, it is the protein that provides
tensile strength. There are 28 different types of collagen assembled from 41 different polypeptidic chains genetically distinct. The collagen is produced by the cells by exocytosis and it is located in the extra-‐cellular matrix and interact with cells via receptors and regulate their proliferation, migration and differentiation.
Collagen is considered as an elastic protein with a resilience of 90%: collagen fibers are indeed able to deform reversibly and their mechanical properties can be investigated by force spectroscopy (26)
The collagen turbidity can be an obstacle to 3D patterning features under tens of micrometers below the surface (31); therefore Applegate et al. (32) used an elastometric silk fibroin hydrogel as a biomaterial. Laser micropatterning of the gel allows the generation of scaffolds with interconnected porous networks with voids of 5 µm in diameter until almost 1 cm below the gel surface by multi photon absorption (MPA). Within a material that is transparent to the low-‐energy photons, very little light is absorbed at the surface which allows to have a focal spot formed and the MPA to occur deep within the material. The laser used was a 810nm with a impulsion time of 100 fs and a repetition rate of 80 MHz, giving a relatively low energy (sub 2 nJ per pulse), with a 10x objective (numerical aperture NA=0.3).
20
The pattern created was a line made by a single pass of the laser at a constant speed of 50 mm/s with varying pulse energies. The minimum pulse energy necessary to observe structural changes in the silk gel is 0.25 nJ per pulse which corresponds to 0.7 𝑚𝐽/𝑐𝑚!.
Yaoming Liu et al. (33) also 3D-‐micropatterned collagen scaffolds by femtosecond laser ablation. A 800 nm, 45 fs Ti:Sapphire laser was used with a 10x (NA=0.25) objective to create various 3D patterns in a collagen gel. The threshold fluence for ablation of the scaffold was found to be 0.06 J/cm2. Mesenchymal stem cells from rat bone marrow and human fibroblasts were then seeded within the ablated patterns and were shown to be viable for at least 10 days.
Nanorod particles can also be used in order to 3D cell patterning in collagen I hydrogel (34). Gold nanorods are added to the collagen hydrogen used in order to increase the near-‐infrared femtosecond laser beam. They absorb the NIR light (800 nm, 100 fs and 90 MHz) and release the energy in the form of photons, using photothermal effect, which thermally denaturates the surrounding collagen matrix. Channels are patterned inside the hydrogel with a laser objective lens (10X, NA 0.45). The pattern resolution can be tuned by adjusting the laser power and the writing speed of the laser. For a high pattern fidelity and a high cell viability (> 90%), a writing speed of 2.0 mm/s and a laser power of 100 mW were used that is a fluence of 54 𝐽/𝑐𝑚!. An inverse relationship exists between the writing speed and pattern width. The laser power determines the pattern size. These channels produced within the hydrogel cause cell migration and 3D cell alignment to the patterns over a period of 14 days.
D. The project’s objectives
This project was led in collaboration between ALPhANOV, a technologic center for optics and lasers (in Talence, France) and the BioTis INSERM (National Institute for Health and Medical Research) U1026 Unit (Tissue Engineering Unit, in Bordeaux, France). The biological part of the project was carried out in the INSERM labs and all the optics and photonics parts in APhANOV.
BioTis U1026 INSERM lab is amongst others focusing on bone tissue regeneration. In order to overcome the vascularization issue and prevascularize bone tissues engineered at the lab, an easy-‐to-‐use and versatile method for building cell-‐laden microfibers has been developed (38). The objective of the project was to perfuse these microfibers in order to use them as substitutes for capillaries. First tests performed on these microfibers showed that they cannot be perfused if not pre-‐holed.
The main idea is to pre-‐pattern microfibers with a laser-‐based approach. The achievement of void creations inside the microfibers would allow an easier perfusion without cell damage.
Some preliminary trials to create a lumen into the microfibers had been performed using a similar laser setup in 2014 and showed the potential of the approach.
21
The aim of this final degree project was therefore to demonstrate the feasibility of the process and to optimize this process.
II. Material and methods
A. Preparation of the fibers
1. Cell culture D1 cell line was purchased from ATCC (LGC Standards, Molsheim, France). These cells come from rat bone marrow. The D1 cell line was cultured in 150 cm2 culture flasks, with Dulbecco's Modified Eagle Medium (DMEM, Gibco, Life Technologies) supplemented with 10 % fetal calf serum (FCS) (GE Healthcare®) and 1/1000th plasmacin (plasmacinTM prophylactic, 2.5 mg/mL, Invivo) at 37°C and with 5 % CO2. Endothelial cells were cultured in 150 cm2 culture flasks, with Dulbecco's Modified Eagle Medium (DMEM, Gibco, Life Technologies) supplemented with 10 % fetal calf serum (FCS) (GE Healthcare®) and 1/1000th plasmacin (plasmacinTM prophylactic, 2.5 mg/mL, Invivo) at 37°C and with 5 % CO2.
They were mixed with neutralized collagen in order to obtain a final cell density of 50 x 106
cells/mL on the fiber.
The confluent cells were collected from the culture flask by a trypsin treatment (trypsin solution at 0.1 mg/mL in EDTA 0.065 mg/mL).
2. Collagen
The collagen used was Type I Collagen purchased from BD Biosciences (Collagen Type I, obtained from rat tail, Bedford, US). The type I collagen used comes from rat tail tendons and its quality depends on the arriving lots, the age and health of rats inducing variable deterioration.
3. Fiber manufacturing
The making of the fibers was done under sterile condition, on a microbiological safety workbench, according to the protocols adapted by Kalisky et al (38). The general idea was to create a shell around a glass capillary and then replace the glass capillary inside the sheath by the solution of collagen to create the fiber.
The sheath is made by cross-‐linking sodium alginate (10/60 at 2 % in PBS 1X (Phosphate Buffer Saline)) with calcium chloride (CaCl2) solution at 0.2 mol/L. Sodium alginate Protanal LF-‐10/60 was purchased from (FMC Biopolymer, Drammen, Norway and calcium chloride was purchased from Sigma (Saint Quentin Fallavier, France) and PBS 0.1 M pH 7 was from Gibco (Life Technology SAS, Saint Aubin, France). A 300 µm diameter glass
22
capillary is plunged into calcium chloride and alginate starting and finishing with CaCl2 (see Fig 9 A). The sheaths are left in the cell culture medium.
In the meantime, after having trypsinated the cells and collected them in the bottom of the tube, the cell/collagen solution is prepared. All the preparation takes place in the ice so that the collagen does not jellify. The collagen solution is composed of:
-‐ 𝑉! of collagen (𝐶1 ∗ 𝑉! = 𝐶2 ∗ 𝑉!"!#$ with 𝐶1 = 8.38 𝑚𝑔/𝑚𝐿 the stock solution collagen concentration, 𝐶2 = 5 𝑚𝑔/𝑚𝐿 the final collagen concentration, V2 the total final volume of the solution).
-‐ 𝑉!"#$ = 𝑉! ∗ 0.023 of sodium hydroxide (5 𝑚𝑜𝑙/𝐿, to neutralize the pH of Collagen from 2 to 7)
-‐ 𝑉!"#!"! = 𝑉!"!#$/10 of 10xPBS solution -‐ 𝑉!"!"#$ = 𝑉!"!#$-‐ 𝑉!-‐𝑉!"#$-‐𝑉!"#!"!
The concentration needed in the final solution to be able to seed the fibers with cells is 50 million cells/mL. From this concentration, the total volume 𝑉!"!#$ to prepare can be deduced, depending on the number of cells collected during trypsination.
This collagen solution is mixed up with the cells with a vortex. It is then added in a syringe that was previously placed in the freezer. The syringe is introduced at the end of the 300 µm diameter glass capillary – inside the alginate sheath: the goal is to replace the capillary by the solution by removing progressively the glass capillary while keeping the syringe triggered (see Fig 9 B).
Figure 9: A) Creation of the sheath around the glass capillary B) Replacement of the glass capillary by the
collagen/cell solution (38)
The whole fibers and sheath are placed in the incubator at 37°C during 12 to 24 hours; the fibers will jellify and the cells will migrate to the interface collagen-‐alginate. The diameter of the created fibers varies but is usually between 100 and 150 µm. The fibers are smaller than the 300 µm diameter glass capillary used to create the sheath because when placed in the incubator for 24 h, the collagen jellified and the cells got closer and tighter, retracting the fiber in length and diameter. The longer the fiber is left in the incubator, the more contracted it will be.
23
4. Fixation of fibers
Some of the used fibers were fixed in order not to have the viability condition on the first tests performed on the D1 cell fibers.
The cells were fixed with a 4 % paraformaldehyde (PFA) solution at 4°C for 30 minutes. They were then rinsed out three times with a PBS 1X solution.
To unsheathe the fibers two processes can be used. The first one is unsheathing manually: dismantling of the sheath with the tweezers and a scalpel and with care. The second one is plunging the fibers in a sodium citrate solution at 55 𝑚𝑜𝑙/𝐿 kept at 37°C to accelerate the reaction: the sodium atoms take the place of the calcium atoms and un-‐crosslink the sheath. A representation of the fiber before and after the unsheathing can be seen on Fig 10. The D1 fixed fibers are then kept in distilled water at 4°C for several weeks.
Figure 5: Fiber with (on the left) and without (on the right) the sheath. The green part is the collagen, on the edge there
are the cells (38)
B. Micromachining the fibers
1. The laser sources Lasers have been used for the experiments, in order to study the influence of different pulse duration and different wavelengths.
a. Sirius
The Sirius laser is a laser developed by Spark Lasers (Talence, France). It delivers 10ps pulses at 1064 nm with a repetition rate varying from 20 kHz to 2 MHz. At 50 kHz, its maximum energy delivered is above 60 µJ per pulse. The average power at the laser output is 3 W. A frequency doubling system is added at the output of the laser, allowing to choose the wavelength between 1064 nm and 532 nm. This laser was used at 1064 nm and 532 nm with a repetition rate of 50 kHz.
b. Satsuma/eclipse II
The Satsuma laser is an Ytterbium-‐doped fiber laser developed by Amplitude Systems (Talence, France). It delivers 350 fs pulses at 1030 nm with a repetition rate of 500 kHz to 2 MHz. Its maximum energy delivered is 10 µJ per pulse. At 500 kHz, the average power at the laser output is 5 W. A frequency doubling system is added at the output of the laser, allowing to choose the wavelength between 1030 nm, 515 nm and 343 nm. This laser was used at 1030 nm with a repetition rate of 50 kHz.
24
2. The optical set ups
a. The optical set up Eclipse 1
The Eclipse 1 set up is a bottom-‐up machining system. It was adapted to several laser sources during this project. A half-‐wave plate and a polarized beam splitter cube were inserted in the optical path at the laser output. This allowed an extra laser cavity tuning of the average power of the laser. The laser beam was inserted in an inverted microscope (NIKON Eclipse Ti-‐U) and is focalized by a Mitutotyo objective (magnification x20/NA=0.4) (see Fig 11 & 12). In order to cover totally the entrance pupil of the objective (8 mm), a beamexpander is added in the optical path, according to the size of the laser beam at the output of the laser and the distance between the laser output and the entrance of the inverted microscope.
A CCD camera allows to visualize the sample via transmission lighting from under the objective (see Fig 14).
The laser focal spot is fixed while the sample is moving thanks to a system of XYZ motorized stages (AEROTECH, A3200). Theses stages were controlled by a G-‐code program (Annex 1). The resolution of this stages system is 250 nm and the maximum speed that can be reached in this configuration is 10 mm/s. The pattern created was a line made by a single pass of the laser at a constant speed of 0,5 mm/s.
After aligning the set up on the Sirius laser (infrared wavelength (IR)), the laser beam was naturally diverging enough in order to cover the pupil of the objective (the beam had a much bigger diameter than 8 mm, rather 20 mm).
When the set up was mounted on Sirius laser (green wavelength), a beam expander was used in order to cover totally the pupil of the objective. The alignment of the optical path was finely tuned using the autocorrelation method.
Figure 6: Eclipse I inverted microscope, Aerotech stages and sample
25
Figure 7: Scheme of the Eclispse 1 set-‐up
b. The optical set up Eclipse II
The Eclipse 2 set-‐up was already in place and aligned. It is a machining system allowing machining from the top. The Satsuma laser was used with this machining set up. As for the Eclipse 1 set up, a half-‐wave plate and a polarized beam splitter cube were inserted in the optical path at the laser output, as well as a beam expander. This allows an extra laser cavity tuning of the average power of the laser. The sample was mounted on stages, themselves mounted on a granit gantry that allows stability. The laser beam was inserted in one of the gantry openings and goes through a Mitutotyo objective (magnification x20/NA=0.4) (see Fig 13 & 14). In order to cover totally the entrance pupil of the objective (8 mm), a beam expander is added in the optical path.
A CCD camera allows to visualize the sample via transmission lighting from above the set up. (see Fig 13)
The sample is moving thanks to a system of motorized XYZ stages on an hexapod (prototype created by ALIO, with 5 axes – X, Y, Z, and two rotational axes). Theses stages are controlled by a G-‐code program. The resolution of this stage system is 1 µm and the maximum speed that can be obtained is 100 mm/s.
Figure 8: Scheme of the Eclipse II set-‐up
26
Figure 9: Photo of the Eclipse II set-‐up
c. Summary of the characteristics used for the different tests Here is a table (Table 1) that sums up the characteristics of the different performed tests. Table 1: Characteristics of the different tests performed
Wavelength Repetition rate (used)
Energy (used)
Pulse duration
Theoretical spot diameter
Theoretical fluence
Sirius (IR) – Eclipse 1
1064 nm 50 kHz 3 µJ 10 ps 3,2 µm 36,2 J/cm2
Sirius (green)
(Annex 2) – Eclipse 1
532 nm 50 kHz 3 µJ 10 ps 1,6 µm 145,1 J/cm2
Satsuma (IR) –
Eclipse 2
1030 nm 50 kHz 2.92 µJ 350 fs 3,1 µm 37,7 J/cm2
3. Sample-‐holding systems
Several versions of the sample-‐holding system were developed during the project.
a. Alginate sheath (V1)
The first sample-‐holding system (V1) created was a glass side on which the fiber was laid. The alginate was cross-‐linked with calcium chloride to hold the fiber still. This first configuration has proved that the fiber needed to be laid on a plane sample-‐holding system in order to facilitate the stage movement.
b. Agarose gel with V-‐groove (V2)
The second version of the sample-‐holding system developed (V2) consisted of a 6-‐well plate in which a layer of agarose gel was poured. In this agarose layer, a thin straight V-‐
27
groove had been formed by putting two microscope plates glued together inside the gel while it was solidificating. In this V-‐groove, the fiber was positioned with caution. This set up allows to keep the fiber hydrated (with the agarose gel around). Besides, it enables to keep the fiber straight without movement during the machining.
Nevertheless, the drawback of this set up is that it is not convenient for transportation between the two working places (ALPhANOV and INSERM) as most of the fibers got lost during the transportation. Unfortunately, it introduces spherical aberration for bottom-‐up machining (See Annex 2).
c. Airtight enclosure (V3) The third version of the sample-‐holding system (V3) was a closed small box with a
glass plate bottom. The box was kept closed with a humid tissue inside in order to keep the fiber moisturized. This 3rd version allowed to get rid of the spherical aberrations. The limit of this set up is that the lid has to be removed to machine the fiber on top-‐down machining (Eclipse 2 set up). This not only decreases the sterility of the experiment but also decreases the available time for the experiment as the fiber dries out quicker. This is, however, still reasonable as a fiber dries out in about 45 minutes, which leaves a sufficient amount of time to process it.
C. Characterization method: confocal fluorescent microscopy
Confocal fluorescent microscopy was used in order to visualize and characterize the voids created by laser machining but also to test the viability of the cells after machining thanks to fluorescent labels. Confocal microscopy allows to get increased optical resolution and contrast.
1. DAPI Labeling The nuclei of the cells from the fixed fibers were labeled via a DNA-‐labeling with DAPI (diluted at 1/5000 in PBS 1X). The fibers were left 10 min at ambient temperature in total darkness in the DAPI solution. They were then rinsed with distilled water. A confocal microscope (Leica TCS SPE, Model DMI 4000B) has been used to highlight whether the fibers were machined or not.
2. Live/Dead® Assay
The cells of the non-‐fixed fibers were labeled with a Live/Dead® assay (Thermofisher scientific) four hours after the machining. This allows a convenient discrimination between live and dead cells. In a relatively low brightness environment, the Live-‐Dye™ (a cell-‐permeable green fluorescent dye (Ex/Em = 488/518 nm) stains live cells), and the Dead-‐Dye (ethidium homodimer-‐1 a cell non-‐permeable red fluorescent dye (Ex/Em = 488/615) stains the dead cells) solutions from the Live-‐Dead Staining Kit were added to 2.5 mL of culture medium. This solution was then placed at 37°C and 5 % CO2 in an incubator for 10 min. The culture medium in which the fibers were plunged was withdrawn, in the meantime, and the solution was added to the culture well in which the fibers were. The fibers were left in the stove at 37°C and 5 % CO2 for 15 min. The solution was then taken out of the well and the
28
fibers were rinsed with cell culture medium. The fibers were then observed under the confocal microscope (Leica TCS SPE, DMI 4000B model).
III. Results
A. Influence of the laser parameters on the machining
1. Influence of the energy: threshold determination
The influence of the energy was studied in two different set ups: on the Eclipse 1 set up mounted on the Sirius laser (IR, picosecond regime) and on Eclipse 2 set up mounted on the Satsuma laser (IR, femtosecond regime). The goal was to determine the threshold energy from which a precise machining is possible in the fiber without damaging the surrounding fiber (splitting it or tearing it in several parts).
The third version of sample-‐holding system (V3) was used for these tests.
a. Tests on Sirius (IR)/Eclipse 1 set up
The characteristics used with the Sirius laser (IR) on fixed D1 cell-‐laden fibers are summed up in Table 2.
Table 2: Parameters used in Sirius (IR) tests
Wavelength Repetition rate
Energy per pulse
Pulse duration
Cells used
Sirius (IR) 1064 nm 50 kHz 2.2 µJ and 3 µJ
10 ps Fixed D1
The main energy tests are summurized in Table 3:
Table 3: Energy tests on Sirius (IR)
𝑷𝒕𝒂𝒓𝒈𝒆𝒕 (mW) Energy (𝛍𝐉) Number of fibers Result: did it work? 50 1 7 No (n=7) 108 2.2 4 No (n=2)
Yes( n=2) 150 3 10 Yes (n=9)
No (n=1) A ‘yes’ result in this table corresponds to an observable channel created inside the
fiber, as in Fig 15 or Fig 16. At 2.2 µμJ, four fibers were machined, and 50 % of these presented observable channels during confocal observation (see Fig 15). At 3 µμJ, ten fibers were machined and 90 % of these presented observable channels during confocal observation (see Fig 16)
29
The size of the channels were measured with the tool of the confocal software. The fibers machined at 2.2 µμJ showed an average channel diameter of 8.65 𝜇𝑚 wide (see Table 4). The fibers machined at 3 µμJ had an averaged channel diameter of 10.07 𝜇𝑚.
Figure 10: Confocal image of a fiber machined at 2.2 µJ: three views (a) DAPI labelling in the cells (b) autofluorescence of
the collagen (c) merge of the two images (cells and collagen)
Figure 11: Confocal image of the fiber machined at 3 µJ (merge of the two images)
Table 4: Size of the channels depending on the energy for a machining with Sirius laser (picosecond).
Number of measures
Number of fibers processed
Average diameter of the canal (𝜇𝑚)
Standard deviation of the diameter of the canal (𝜇𝑚)
At 2.2 𝜇𝐽 119 2 8.65 3.11 At 3 𝜇𝐽 71 2 10.07 3.63
b. Tests on Satsuma (IR)/Eclipse 2 set up
The tests performed with Satsuma laser (IR) on fixed D1 cell-‐laden fibers are summarized in Table 5.
Table 5: Parameters used in Satsuma (IR) tests
Wavelength Repetition rate
Energy per pulse
Pulse duration
Cells used
Satsuma (IR) 1030 nm 50 kHz 3 µJ 350 fs Fixed D1
30
The main energy tests performed on Satsuma (IR) are summarized Table 6.
Table 6: Energy tests on Satsuma (IR)
𝑷𝒕𝒂𝒓𝒈𝒆𝒕 (mW) Energy (𝛍𝐉) Number of fibers Result: did it work? 150 3 20 Yes n=18
No n=2 A ‘yes’ result in the table corresponds to an observable channel created inside the
fiber, as in Fig 17 to 19. At 3 µμJ, twenty fibers were machined, and 90 % of these presented observable channels during confocal observation. The created channels were well distinguishable on the confocal observation. The average diameter of the channels created inside the fibers 12.8± 4.5 µμ𝑚 (average of 78 values of the diameter from three different fibers patterned). It was measured with the measure tool of the confocal software.
Figure 12: Zoom on a part of the fiber (a) DAPI label on D1 cells (b) autofluorescence of the collagen (c) merge of the two previous images
Figure 13: Zoom on a part of the fiber (a) DAPI label on D1 cells (b) autofluorescence of the collagen (c) merge
of the two previous images
31
Figure 19: Zoom on a part of the fiber (a) DAPI label on D1 cells (b) autofluorescence of the collagen (c) merge
of the two previous images
2. Influence of the pulse duration on the machining
The influence of the pulse duration was studied by comparing two different pulse duration: the picosecond regime, with Sirius laser on the Eclipse 1 set up and the femtosecond regime with Satsuma laser on Eclipse 2 set up at comparable energies (3 𝜇𝐽 in both regimes). The size of the created channels in the fibers are summarized in Table 7. Table 7: Size of the channels depending on the pulse duration for a machining at 3 µJ
Number of measures performed
Average diameter of the canal (𝜇𝑚)
Standard deviation of the diameter of the canal (𝜇𝑚)
Picosecond regime
71 10.07 3.63
Femtosecond regime
78 12.79 4.48
3. Influence of the wavelength on the machining
Some tests were performed in order to study the effect of the wavelength on the machining. The settings of the lasers used are summarized in Table 8. Table 8: Characteristics of the lasers used for wavelength-‐dependent test
Laser used Wavelength Repetition rate
Energy per pulse
Pulse duration Cells used
Satsuma (green)
515 nm 50 kHz 2.12 µJ 350 fs Fixed D1
Sirius (green) 532 nm 50 kHz 3 µJ 10 ps Fixed D1
Some tests were performed with the frequency doubling systems on Satsuma (green) laser with Eclipse 2 set up and Sirius (green) laser with Eclipse 1 set up.
32
Nevertheless, no machining could be observed on the fibers machined with Satsuma (green) laser and Eclipse 2 set up under confocal microscopy. The results on Sirius (green) with Eclipse 1 set up highlighted the limits of the sample-‐holding system V2 (see Annex 2).
B. Fabrication of capillary substitutes
1. Whole fiber machining Satsuma IR with Eclipse 2 set up was used in order to test whole fiber machining as they present optimal conditions. The mean target power was 146 mW. Whole fixed fibers and unfixed fibers were processed, using a bit-‐by-‐bit machining: as the fiber is very inhomogeneous, in diameter, the processing of the fiber was done only on a few hundreds of micrometers each time, adjusting the position of the fiber after each shoot thanks to stages.
The machining was achieved on ten whole fibers with a length of 1.5-‐2 mm (Fig. 20). Each fiber was machined in approximately 20 to 25 minutes.
Figure 14: Whole fiber machined by Satsuma laser (IR)
2. Cell viability
In order to test the cell viability during the process, tests were carried on non-‐fixed fibers, laden with endothelial cells. These tests were performed on the Sirius (IR) machining tests on Eclipse 1 set up.
Live/dead assays were performed after the patterning in order to check for cell damage from the exposure of the layer of cells to the laser beam (Figs 21 and 22).
33
Figure 15: Live/dead protocol on a non-‐fixed fiber of endothelial cells machined at 3 µJ: observation at the confocal microscope (green: live cells, red: dead cells). This live/dead assay is representative for n=6 fibers
On Figure 22, the two top images show respectively the dead cells (in red) and the live cells (in green). Below, a merge of the two top images give three different sections (2 longitudinal sections along different axes, and a transverse section). Most of the cells were viable except in one place: this place corresponds to the place where the laser shot through the cell layer, as we can see on the transverse section.
Figure 16: Live/dead assay on a non-‐fixed fiber of endothelial cells machined at 3 µJ: observation on the
confocal microscope (green: live cells, red: dead cells), on thee different sections (longitudinal and transverse sections)
IV. Discussion A. Influence of the laser parameters on the machining
1. Influence of the energy on the machining: threshold determination The tests on the threshold energy showed that under 2 µμJ, no patterning is observed. At
2.2 µμJ, 50 % of fibers were patterned, and at 3 µJ, 90 % were patterned. Therefore, the threshold energy seems to be around 2.2 µμJ, but the threshold energy for a reproducible
34
machining of the fibers is 3 µJ. This corresponds to a fluence of 36.2 J/cm2 on Sirius Laser and 37.7 J/cm2 on Satsuma laser. The non reproducibility of the machining at 2.2 µμJ could be due to the inhomogeneity of collagen: indeed, depending on the microfiber machined, a different pattern can be obtained. The collagen used is extracted from rat bone marrow. The health and age of the rats used vary inducing a huge variability in the collagen used to create the microfibers. Therefore, the conclusion is that a minimum energy of 3 µμJ should be used to pattern the fibers. However, the channels created had a quite similar diameter with an energy per pulse of 2.2 µμJ and 3 µμJ: 𝑑!,!!! = 8.65± 3.11 µμm and 𝑑!!! = 10.07± 3.63 µμm. Therefore, there is a relative flexibility in the energy that has to be used on the fibers. The sample-‐holding system (V3) induces no spherical aberration in the laser beam, which creates a focal spot size close to the theoretical size. As the fluence is evaluated as the target energy divided by the surface of the spot, having a tiny focal spot allows to maximize the fluence for a fixed energy. Therefore, a smaller energy is needed to get the same fluence. This way the collateral damages on the cells can be reduced.
Applegate et al. (32) were also using a femtosecond laser with an pulse duration close to ours (100 fs), but with a slightly different wavelength (810 nm) and a numerical aperture of 0.3. The difference between his approach and the one developed here is that the voids Applegate created were patterned in a hydrogel block that was seeded with cells. The cell-‐laden fibers were patterned after the cells were seeded on the fiber. Applegate was obtaining thinner voids (5 µm in diameter). This is probably due to the difference in the used energies (under 2 nJ per pulse for Applegate versus 3 µJ for our study), and the difference of repetition rates (80 MHz versus 50 kHz).
2. Influence of the pulse duration Yaoming Liu et al. (33) used a similar wavelength (800 nm) but with a much smaller pulse duration (45 fs) and a smaller numerical aperture (NA=0.25). Their threshold fluence for ablation of the scaffold was 0.06 J/cm2 which is also much smaller than what was obtained here because the bigger pulse duration might have produced thermal effects on the collagen in our case.
At 3 µμJ (fixed energy), the standard deviation of the results in femtosecond regime (fs)
and picosecond regime (ps) overlap. Therefore, it seems that the pulse duration does not have an impact on the channel diameter. Given that femtosecond lasers are much more expensive compared to picosecond lasers, this is a relevant information to decrease the general cost of the process. By using a picosecond laser, the patterning of collagen microfibers would therefore be accessible at much lower costs, which, in a purpose of future commercialization of the process is a huge advantage. Nevertheless, further tests should be done using Second Harmonic Generation (SHG) in order to study the impact of fs versus ps on the collagen fibrils inside the fibers. Indeed, femtosecond lasers do not have thermal effects on matter, but the thermal effects usually
35
occur from a few tens of picoseconds which is the pulse duration of Sirius laser. Therefore, there might be some thermal effects affecting the collagen fibrils of the fiber. A SHG imaging would allow to check the orientation of the fibrils and check how the lasers modify the collagen hydrogel.
3. Influence of the wavelength
The tests performed at 532 nm were not conclusive. Indeeed, the sample holding system V2 does not suit for green laser light machining on Eclipse 1 because the energy needed to engrave the glass (0.4 µμJ) is lower than the energy needed to engrave the microfiber of collagen (3 µμJ).
Besides, the absorption coefficient of the water has a zero value at 532 nm (Fig 23). The collagen fibrils absorption coefficient has an order of magnitude of 102 cm-‐1. The collagen hydrogel used to build the fibers is composed of 95 % of water so the collagen fibril absorption coefficient has no influence on the machining. This is an explanation of why the fibers processed at the green wavelength did not have any channel created.
Figure 17: Absorption coefficients of several species including collagen (collagen fibrils) and water (39) Hribar (34) added gold nanorods to the collagen hydrogen used in order to increase
the near-‐infrared femtosecond laser beam absorption to thermally denaturates the surrounding collagen matrix with a low fluence machining. However, we did not use any gold nanorod particles in order to use as little external elements as possible from the natural environment of the cell. Nevertheless, the approach developed in our project has a fluence smaller than the one used by Hribar (fluence of 54 𝐽/𝑐𝑚! for him, fluence of 37 𝐽/𝑐𝑚!).
B. Fabrication of blood capillaries substitutes
36
1. Machining on the collagen
On all the confocal images, a lack of collagen is observed (black regions in the green autofluorescent collagen). A change of contrast or apparition of bubbles were also observed on the CCD camera during the machining. The void created pushed collagen on the sides of the void. As the collagen is a hydrogel (95 % of water), the water filled the voids created by the laser machining inside the fiber. As water and collagen have a similar refraction index, the contrast between the collagen and the void disappeared rapidly (within a few minutes maximum) when observed on the CCD camera. The sample had to be observed under confocal microscopy to visualize the collagen-‐free region inside the fiber.
2. Size of the channels
The theoretical spot size of the laser beam at the focal point was 3 µm. The huge difference between the theoretical spot size and the void created can be explained by the high frequency rate of the laser (50 kHz): the optical breakdown zones are created close to each over, even overlapping. Besides, if a void is created directly where the laser is focused, the water contained in the collagen fills the void after the machining and can play a magnification effect when the laser beam goes through the collagen, focusing on a different place than originally planned. At the start and stop places of the laser shoot, a bigger void can be observed (20 to 30 µm) (Fig 20 and 21). This could be due to the accelerating and decelerating phases of the stage motion. In the channels created, some parts of the voids are larger than others (up to 20 or 30 µm wide -‐ see Figs 20 and 21). This corresponds to the accelerating and decelerating phases of the stages motion: it slows down the pace of the machining and increases the overlap of the focal zones between two shots; therefore creates bigger voices.
3. Whole fiber machining: choice of optimal settings
Satsuma IR with Eclipse 2 seems to offer the best machining set up. The Eclipse 2 allows machining from the top which prevents the laser beam to cross a medium different than air when patterning the fiber.
Besides, it offers a good visualization of the fiber and easy focalization of the laser
beam in the middle of the fiber. The tests done on the Eclipse 1 set up have all faced the same issue with focalizing the laser beam in the middle of the fiber: in bottom-‐up machining, the CCD camera does not allow to see the top of the fiber but only the glass plate on which the fiber lies. On Eclipse 2 set up, the CCD camera is above the sample which means both the top of the fiber and the glass plate of the sample-‐holding system are sharp and easily observed. Therefore, the average between two ‘z’ positions (top and bottom of the fiber) can easily be found with the stages and the middle of the fiber assumed. Eclipse 2 gives the best results in terms of repeatability and diameter control of the voids.
37
The optimal parameters that give a repeatable pattern on the fibers are summed up in Table 9. Table 9: Optimal parameters
Set up Eclipse 2 + Satsuma laser Wavelength 1030 nm Pulse duration 350 fs Repetition rate 50 kHz Mean power 150 mW Energy!"#$%! 3 µJ Stage speed 0,5 mm/s
This is the Eclipse 2 set up that influences the machining, so another laser than Satsuma laser could be used.
In order to achieve a clean patterning, the machining should not be done on more than 300 µμm in a row. Indeed, the fibers do not have a smooth surface (they can be 50 µm thick in one place and 100 µm thick 300 of µm further in the length of the fiber). Therefore, the machining has to be adapted that is shooting slantwise into the fiber, without shooting the cells on the outskirt.
4. Cell viability The tests on unfixed cell-‐laden fibers gave interesting results concerning the viability of the cells: the live/dead protocols showed that most cells are alive after the transport and machining process, except where the laser got through the cell layer through the cell layer. This is encouraging for further developments of the process even if quantification essays and DNA analysis have to be performed to assess cells viability.
Conclusion and Perspectives Micromachining of the fibers was achieved with femtosecond and picosecond lasers,
demonstrating the feasibility of the process. The energy threshold for laser machining was 3 µJ, whatever pulse duration was used.
It was found that there is a great freedom on the parameters for machining: a change
in the target energy of 0.8 µJ only changes the mean average diameter of the channels of 1.42 ± 3 µμ𝑚, that is a 0.8 µJ change of energy does not have any effect, regarding the standard deviation obtained. Besides, the change from picosecond to femtosecond only changed the diameter of the channel of 2.7 ± 4 µμm, which means the pulse duration does not have any effect, regarding the standard deviation on the result.
38
This freedom in the parameters offers a wide range of choices in the lasers used for the patterning, which can decrease the general cost of the process.
Further tests should be performed on collagen in order to determine the effects of
different pulse duration (picosecond and femtosecond regimes) and in order to visualize whether collagen remodeling takes place after the laser machining. Femtosecond lasers create an athermal machining: the impulsion lasts less than the characteristic diffusion time of heat. Nevertheless, the Sirius laser has an impulsion length of 5 to 10 ps, which is the limit for thermal effects. Further investigation should be done in order to determine whether there are thermal effects or not.
Besides, tests will be performed in a few weeks in order to perfuse the patterned microfibers. This will be the opportunity to determine whether the fiber has to be totally machined or only partially, which size of voids is needed for an optimal perfusion, etc.
Finally, some tests on the effect of the wavelength should be carried out in order to determine which wavelength could have a positive effect on the surrounding cells. Indeed, in the litterature, some studies have shown that exposition of cells at low powers at 532 nm gives a benefic proliferation and differentiation effects (40–42).
39
Annex 1: Example of a G-code program Here is an example of a G-‐code program used to control the stages : 'This programm writes one line along Y axis. 'Plate movement : line laser OFF -‐ line laser ON-‐ line laser OFF. dvar $VMAX 'Maximal velocity dvar $NL 'Number of lines in serie dvar $STEP 'Separation between lines dvar $LINE 'Line length M0 ENABLE X Y Z $VMAX=0.5 $STEP=5 $LINE=0.8 G92 X0 Y0 Z0 G91 RAMP RATE 1000 VELOCITY ON $AO[1].X = 5 G1 Y -‐$LINE $AO[1].X = 0 G1 Y $LINE VELOCITY OFF G90 G1 X0 Y0 F $VMAX
40
Annex 2: Tests with Sirius laser (green) on Eclipse 1: optimization of patterning set-‐up
A. From the sample holding system V2 to V3
Fixed fibers laden with D1 cells were used for these experiments. The sample holding system V2 was used. The settings of the laser used are summed up in Table 10. Table 1: Settings of Sirius (green) laser for tests on wavelength effect
Wavelength Repetition rate (used)
Energy (used)
Pulse duration
Cells used
Sirius (green) 532 nm 50 kHz 3 µJ (at the objective output)
10 ps Fixed D1 and live D1
The first objective was to test different energies in order to find out the threshold
energy for micromachining the fibers at 532 nm. Target powers from 10 to 500 mW were used (results summed up in Table 11). Table 2: Energy tests on Sirius (green) laser
𝑷𝒕𝒂𝒓𝒈𝒆𝒕 (mW) Energy (𝝁𝑱) Result: did it work? 14 mW 0,2 µμJ No (n=2) 103 mW 2 µμJ No (n=2) 170 mW 3,4 µμJ Fiber exploded (n=1) 224 mW 4,5 µμJ Works out (n=1) 480 mW 8 µμJ Fiber destroyed (n=1)
According to these tests, the threshold power seems to be between 2 µμJ and 3,4 µμJ but the fourth result (at 4,5 µμJ) differs. When the laser is focused in the middle of the fiber, some bubbles appear under the fiber, at the interface with the sample-‐holding system (Fig 24).
Figure 18: Microscope observation of the fiber machined at 4,48 µJ (obj 10x) (a) before (b) after the machining occured
Under confocal microscopy, a void is observable inside the fiber (see Fig 25), with a diameter of 40 µm, and a thin layer of collagen is left on both sides of the hole which means that the fiber was not torn or the cell layer was not damaged.
41
Figure 19: Confocal observation of the fiber machined at 4,48 µJ (three different sections)
A series of tests (on N=4 fibers) were done at 112 mW (2,24 µJ) in order to study the repeatability of the machining. Different types of behaviour were obtained: a fiber was micropatterned (Figure 26A), another was torn (Figure 26C), one was exploded (Figure 26B), and on the last one, some bubbles of void were created (Figure 26D).
Figure 20: Different fibers machined at 2,24 µJ (a) Fiber machined (b) Exploded fiber (c) Torn fiber (d) Void bubbles
appearance in the fiber
The theoretical fluence of this machining process was calculated: the theoretical spot has a radius twice smaller than in the infrared wavelength so 𝑟 = 0,8 µμ𝑚, which gives a theoretical fluence of 145,1 𝐽/𝑐𝑚!, which is quite surprising compared to what had been obtained in the trials with the other lasers (fluence of about 35 𝐽/𝑐𝑚!.
The fluence obtained here is actually the fluence at the output of the microscope objective: the fluence on target is smaller because of the aberrations introduced in the laser beam at the objective output.
Given the difference of fluence, the point spread function with a layer of 400 µm of agarose and without the agarose gel layer have been plotted in order to study the effect of spherical aberration (Fig. 27) (43). The simulation was done under Zemax software (Zemax 13 release 2 -‐ 2014), then the graphs plotted on Microsoft Excel 2011.
42
Figure 21: Point Spread Function of the beam with an agarose gen layer of 400 µm (a) and without (b) the agarose gel
layer
The PSF is much more degraded with an agarose gel layer of 400 µm than without because of the spherical aberrations introduced in the laser beam. The sample-‐holding system was changed to V3 after these simulations.
B. Highlighting the need of Eclipse 2 set-‐up
Fixed fibers laden with D1 cells were used. The following tests were done with Eclipse 1 set-‐up and the sample holding system V3. The settings of the Sirius laser (green) used are summed up in Table 11. N=10 tests were performed on the fibers. Each time, the laser was actually carving the glass plate under the fiber, despite the settings of the focalization of the laser and the camera plane.
A search for optical breakdown threshold in the glass plate showed that the laser systematically engraves the glass plate instead of the fiber with the green wavelength with these settings. The threshold energy for glass carving was found to be 𝐸 = 0,4 µμ𝐽 (𝑃!"#$%! =20 𝑚𝑊) when the focal spot is on the glass surface.
Still at 3 µJ (id est the threshold energy for fiber patterning), the height of the sample was changed thanks to the stage, so that the focal point of the laser beam is a few tens of micrometers above the glass plate. The glass plate was engraved until the focal point of the laser was lifted 40 µm above the glass plate.
Therefore, the threshold energy for optical breakdown in the glass plate is smaller than the one for the collagen fiber: 𝐸 = 3 µμ𝐽 (𝑃!"�!"# = 150 𝑚𝑊 ). Fewer photons are needed at 532 nm to ablate the glass. Indeed, as each photon has an energy twice at 532 nm compared to 1064 nm, less photons are needed to carve the glass. Thus, the glass is carved even if the beam is focalized in the middle of the fiber.
This makes the Eclipse 1 set-‐up not usable for tests at 532 nm. Therefore a top-‐down
machining set-‐up (Eclipse 2 set-‐up) is preferable for fiber patterning.
43
References
1. Drosse I, Volkmer E, Capanna R, Biase PD, Mutschler W, Schieker M. Tissue engineering for bone defect healing: An update on a multi-‐component approach. Injury. 2008 Sep;39:S9–20. 2. Langer R. I ARTICLES. Science. 1993;260:5110. 3. Blitterswijk CA van, Moroni L, Rouwkema J, Siddappa R, Sohier J. Tissue engineering – an introduction. In: Tissue Engineering [Internet]. Elsevier; 2008 [cited 2016 Jun 1]. p. xii – xxxvi. Available from: http://linkinghub.elsevier.com/retrieve/pii/B9780123708694000240 4. Groll J, Boland T, Blunk T, Burdick JA, Cho D-‐W, Dalton PD, et al. Biofabrication: reappraising the definition of an evolving field. Biofabrication. 2016 Jan 8;8(1):013001. 5. Jain RK, Au P, Tam J, Duda DG, Fukumura D. Engineering vascularized tissue. Nat Biotechnol. 2005 Jul;23(7):821–3. 6. 2. Biomaterials2015.pdf. 7. Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, Tomoda K, et al. Induction of Pluripotent Stem Cells from Adult Human Fibroblasts by Defined Factors. Cell. 2007 Nov;131(5):861–72. 8. Katari R, Peloso A, Zambon JP, Soker S, Stratta RJ, Atala A, et al. Renal Bioengineering with Scaffolds Generated from Human Kidneys. Nephron Exp Nephrol. 2014;126(2):119–119. 9. Novosel EC, Kleinhans C, Kluger PJ. Vascularization is the key challenge in tissue engineering. Adv Drug Deliv Rev. 2011 Apr;63(4-‐5):300–11. 10. 407249a0 (1).pdf. 11. Liu Y, Chan JKY, Teoh S-‐H. Review of vascularised bone tissue-‐engineering strategies with a focus on co-‐culture systems: Vascularised Bone Tissue Engineering using Coculture Systems. J Tissue Eng Regen Med. 2015 Feb;9(2):85–105. 12. Janssen FW, Oostra J, Oorschot A van, van Blitterswijk CA. A perfusion bioreactor system capable of producing clinically relevant volumes of tissue-‐engineered bone: In vivo bone formation showing proof of concept. Biomaterials. 2006 Jan;27(3):315–23. 13. Lalan, B.A. S, Pomerantseva, M.D. I, Vacanti, M.D. JP. Tissue Engineering and Its Potential Impact on Surgery. World J Surg. 2001 Nov 1;25(11):1458–66. 14. Rivron NC, Liu J, Rouwkema J, Boer J de, Blitterswijk van C. Engineering vascularised tissues in vitro. Eur Cell Mater. 2008;15:27–40. 15. Rouwkema J, Rivron NC, van Blitterswijk CA. Vascularization in tissue engineering. Trends Biotechnol. 2008 Aug;26(8):434–41. 16. Clark ER, Clark EL. Microscopic observations on the growth of blood capillaries in the living mammal. Am J Anat. 1939 Mar;64(2):251–301. 17. Risau W, Flamme I. Vasculogenesis. Annu Rev Cell Dev Biol. 1995 Nov;11(1):73–91. 18. Kneser U, Polykandriotis E, Ohnolz J, Heidner K, Grabinger L, Euler S, et al. Engineering of Vascularized Transplantable Bone Tissues: Induction of Axial Vascularization in an Osteoconductive Matrix Using an Arteriovenous Loop. Tissue Eng. 2006 Jul;12(7):1721–31. 19. Tremblay P-‐L, Hudon V, Berthod F, Germain L, Auger FA. Inosculation of Tissue-‐Engineered Capillaries with the Host’s Vasculature in a Reconstructed Skin Transplanted on Mice. Am J Transplant. 2005 May;5(5):1002–10.
44
20. Hassen-‐Khodja R, Kieffer E. Radiotherapy-‐induced supra-‐aortic trunk disease: early and long-‐term results of surgical and endovascular reconstruction. J Vasc Surg. 2004 Aug;40(2):254–61. 21. Chen X, Aledia AS, Ghajar CM, Griffith CK, Putnam AJ, Hughes CCW, et al. Prevascularization of a Fibrin-‐Based Tissue Construct Accelerates the Formation of Functional Anastomosis with Host Vasculature. Tissue Eng Part A. 2009 Jun;15(6):1363–71. 22. Rouwkema J, Rivron NC, van Blitterswijk CA. Vascularization in tissue engineering. Trends Biotechnol. 2008 Aug;26(8):434–41. 23. richardson2001.pdf. 24. Ott HC, Matthiesen TS, Goh S-‐K, Black LD, Kren SM, Netoff TI, et al. Perfusion-‐decellularized matrix: using nature’s platform to engineer a bioartificial heart. Nat Med. 2008 Feb;14(2):213–21. 25. Perez-‐Castillejos R. Replication of the 3D architecture of tissues. Mater Today. 2010 Jan;13(1-‐2):32–41. 26. Ricard-‐Blum S. The Collagen Family. Cold Spring Harb Perspect Biol. 2011 Jan 1;3(1):a004978–a004978. 27. Hoch E, Tovar GEM, Borchers K. Bioprinting of artificial blood vessels: current approaches towards a demanding goal. Eur J Cardiothorac Surg. 2014 Nov 1;46(5):767–78. 28. Engler AJ, Humbert PO, Wehrle-‐Haller B, Weaver VM. Multiscale Modeling of Form and Function. Science. 2009 Apr 10;324(5924):208–12. 29. Murphy SV, Atala A. 3D bioprinting of tissues and organs. Nat Biotechnol. 2014 Aug 5;32(8):773–85. 30. Schiele NR, Corr DT, Huang Y, Raof NA, Xie Y, Chrisey DB. Laser-‐based direct-‐write techniques for cell printing. Biofabrication. 2010 Sep 1;2(3):032001. 31. Smith NI, Fujita K, Nakamura O, Kawata S. Three-‐dimensional subsurface microprocessing of collagen by ultrashort laser pulses. Appl Phys Lett. 2001;78(7):999. 32. Applegate MB, Coburn J, Partlow BP, Moreau JE, Mondia JP, Marelli B, et al. Laser-‐based three-‐dimensional multiscale micropatterning of biocompatible hydrogels for customized tissue engineering scaffolds. Proc Natl Acad Sci. 2015 Sep 29;112(39):12052–7. 33. Liu Y, Sun S, Singha S, Cho MR, Gordon RJ. 3D femtosecond laser patterning of collagen for directed cell attachment. Biomaterials. 2005 Aug;26(22):4597–605. 34. Hribar KC, Meggs K, Liu J, Zhu W, Qu X, Chen S. Three-‐dimensional direct cell patterning in collagen hydrogels with near-‐infrared femtosecond laser. Sci Rep. 2015 Nov 25;5:17203. 35. Niemz MH. Laser-‐Tissue Interactions [Internet]. Berlin, Heidelberg: Springer Berlin Heidelberg; 2007 [cited 2016 Jun 1]. (Biological and Medical Physics, Biomedical Engineering). 36. Mao SS, Qu�r� F, Guizard S, Mao X, Russo RE, Petite G, et al. Dynamics of femtosecond laser interactions with dielectrics. Appl Phys A [Internet]. 2004 Nov [cited 2016 Jun 1];79(7). 37. Noack J, Vogel A. Laser-‐induced plasma formation in water at nanosecond to femtosecond time scales: calculation of thresholds, absorption coefficients, and energy density. IEEE J Quantum Electron. 1999 Aug;35(8):1156–67. 38. Kalisky J, Raso J, Rémy M, Siadous R, Bareille R, Fricain JC, et al. An easy-‐to-‐use and versatile method for building cell-‐laden microfibers. Nature Scientific Reports. under publication. 2016;
45
39. Scholkmann F, Kleiser S, Metz AJ, Zimmermann R, Mata Pavia J, Wolf U, et al. A review on continuous wave functional near-‐infrared spectroscopy and imaging instrumentation and methodology. NeuroImage. 2014 Jan;85:6–27. 40. Poon VK, Huang L, Burd A. Biostimulation of dermal fibroblast by sublethal Q-‐switched Nd:YAG 532nm laser: Collagen remodeling and pigmentation. J Photochem Photobiol B. 2005 Oct;81(1):1–8. 41. Merigo E, Bouvet-‐Gerbettaz S, Boukhechba F, Rocca J-‐P, Fornaini C, Rochet N. Green laser light irradiation enhances differentiation and matrix mineralization of osteogenic cells. J Photochem Photobiol B. 2016 Feb;155:130–6. 42. Anwer AG, Gosnell ME, Perinchery SM, Inglis DW, Goldys EM. Visible 532 nm laser irradiation of human adipose tissue-‐derived stem cells: Effect on proliferation rates, mitochondria membrane potential and autofluorescence. Lasers Surg Med. 2012 Nov;44(9):769–78. 43. Desrus H, Chassagne B, Moizan F, Devillard R, Petit S, Kling R, et al. Effective parameters for film-‐free femtosecond laser assisted bioprinting. Appl Opt. 2016 May 10;55(14):3879.
Related Documents
