Nano-medicine logic design using
a bio-inspred robotic platform and morphological computation
Robotics Research Preparation Assignment: Research Plan
Thursday, 31st March 2017
Edited by Asheesh Sharma
Candidate: 33074, Student: 1562159
University of Bristol United Kingdom
2017
Contents
Motivation 3 Literature review 5 Timeline 7
Motivation
One of the major challenges in medicine have been target drug delivery. Targeted/Smart drug de- livery is a method of medication to promote the concentration of therapeutic drugs to specific or- gans without affecting the functioning of other body systems[1]. For instance, remarkable suc- cess was achieved by using a bacteria as a delivery vehicle. Magnetococcus marinus tends to swim to- wards lower oxygen concentration along the local magnetic fields. Using a programmed magnetic field, the bacteria were directed to tumors (low oxygen). Thus, the bacteria maintained distance from healthy cells (oxygen rich) while moving deep into tumors [2]. Similarly, many other methods have been proposed for delivering medicines to a target tissue but none of them have attracted more attention than the field of nano-medicine. This approach of is based on the novel idea of designing nanoparticles (size, shape, charge, material, coat- ing, activation etc.) which can deliver drugs to target organs without affecting the surrounding tissues. However, there are three main hurdles in the logical design of nanomedicine. Firstly, it is hard to predict the emergent behavior of trillions of nanoparticles interacting at the same time. Sec- ondly, as the population is increased, the desired behavior may not scale accordingly. Thirdly, the design of nanoparticle may not generalize for differ- ent types of tissue, drug and deployment channel.
To overcome these hurdles, researchers have de- vised many targeting strategies. If the circulation time is directly related to the drug’s success (pas- sive targeting), the nanoparticles can be made to prolong the interaction of drug [3]. To make the same nanoparticle specific to the target, it can be equipped with ligands which are only complimen- tary to the receptors (active targeting) of the target tissue (cells) [4]. However, both of these strategies are not sufficient enough to predict the emergent behavior of nanoparticles interacting in a complex tumor environments. In order to build a prediction
1It should also be noted that the hierarchical order can change.
2Bulldozer particles are equipped with ligands which are complementary to the specific receptors of a tissue.
3Satellite particles are loaded with drugs. Each of these particles has a core which is surrounded by drug molecules. 4Localisation means that the particles can change their
state and have reached the target tissue. For instance, bulldozer can deactivate their links to inform satellites to detached their cores (delivering the drug in process).
model, a systems approach can be employed by assuming that the nano-particles are interacting lo- cally. This, in effect, can lead to emergent behavior which is commonly seen in self-organised systems like swarms [5]. With the immediate realization of the link between bio-mimetic nano-particles and swarms, it can also be assumed that the nanopar- ticles have to be hierarchical1and specialized. For instance, a group of nanoparticles (bulldozers2 ) are allowed to penetrate a tumor before another group of particles (satellites3) can deliver the drug. Since the interactions are dependent on the type of tissue and the design variables are physical, the local communication must be mechanical in form. Furthermore, these mechanical “links” also serve the purpose of the computation required for the entire population to localise4when they reach a tumor.
From a bird’s eye view, the entire system (Figure 1) converges on four basic principles which are also shaping the future of bio-inspired design [6].
1. The outcome of internal control is not the only entity which controls the behavior of a system. Any system (not in ideal conditions) is also affected by its morphology which in- cludes sensor placement, the shape of the body and its materialistic properties.
2. Physical constraints on the morphology (DOFs) of a system defines the dynamics
of interaction with the environment.
3. There is a direct link between embodiment and information. Sensory information and body morphology share the regularities of control to cope with the environment.
4. Together, physical constraints, morphol- ogy, and sensory inputs give rise to self- organization resulting in an emergent behav- ior.
3
Drug carrying particle
2
1
1
2
3
Satellite particles are loaded with drugs. Each of these particles have a core which is surrounded by drug molecules. They passively target the tissue (i.e. designed to prolong the interation of drug). When localised, the core detaches itself from the drug molecules, making them active.
Hierarchical and specialised swarm of nano- particles. The swarm consists of bull dozer particles (to penerate inside the tissue) and drug carrying satellite particles (To deliver the drugs). Particles can interact locally using physical means. The mechanical links also provide computation for particles to change their states. For instance, bulldozer particles deactive their links to allow core detachment in satellite particles.
Bulldozer particles are equipped with ligands which are complementary to the specific receptors of a tissue.
Mechanical Links
3
Interestingly, there are many animals which follow these basic principles and resemble the aforemen- tioned swarm like nanoparticle system. Although the study of the swarms has a long history, there is no clear distinction between different types of resulting emergent behaviors. In literature, many definitions are hint on different categories of a swarm.
1. A group of independent, multi-cellular organ- ism cooperating (temporarily) with each other to achieve a certain task. This type of behavior is seen in birds [7] and fishes [8].
2. A group of dependent colonial, multi-cellular organisms cooperating with each other. Such or- ganism often live in colonies and have a fixed spe- cialized. This behaviour is often seen in termites[9], ants[10] and bees [11].
3. A group of dependent or independent, single cell organisms cooperating with each other. There are many species of Bacteria [12] and Myxomycetes[13] which belong to this group.
However, there is a class of organisms which do not belong to any of the groups but still highly resem- ble the functioning of the proposed nanoparticle system. Siphonophores, like many other colonial
hydrozoans in the clade, asexually reproduce a colony of clonal, physically attached, physiologi- cally specialized bodies (called zooids). Due to a high degree of physiological integration, researchers consider these zooids to be dependent but free- living individuals. Each of these zooids is also func- tionally specialized, and show a division of labor
– some zooids are specialized for locomotion, oth- ers for feeding, reproducing, digesting, protecting and so on. More interesting is how Siphonophores flawlessly integrate swarm intelligence with the un- derlying principles of morphological computation, embodiment, and self-organisation.
Due to these reasons, they are the primary inspira- tion for the design of the proposed robotic platform. Using the robotic platform, logical rules will be abstracted which can then be implemented in the functioning of the aforementioned nanoparticle sys- tem. In order to build the robotic platform inspired from Siphonophores, each individual bot will have some specialization and will be constrained to use locally available mechanical information (i.e. each link can be pulled, pushed and stiffened to relay information to the neighbors). The platform is described in the figure.
Bull dozing particle
Figure 1: A simple hybrid nano-particle systems approach
4
RTS control mechanism
Feedback from the four oscillators is equally weighted to dynamically change the stiffness of the spring (spring constant). Thus, changing the resting length of spring allows the robot to retract back from tricky situations
Obstacle avoidance mechanism
Inspired by whiskers, the sensors enable the robot to perceive its environment. The sensor outputs are coupled with mechanosensory oscillators, which drive the RTS control mechanism and the propulsion system. A single robot has four secondary oscillators which take feedback from different groups of whiskers (left, right, front and back). The output of these oscillators is then combined to interact with neighbors and the environment.
Propulsion system
Consists of a thin membrane which triggers pumping action. The frequency of pumping is also controlled by the same group of oscillators.
Real-time tunable spring
It is consists of a DC-motor, winding mechanism and a spring. The resting length of spring is controlled by the combined output of the four oscillators.
Differently specialised robot
Some robots may lack certain features, which may make them reliant on other for certain tasks. For instance, this robot totally relies on RTS and does not have propulsion system
Figure 2: A robotic swarm inspired by Siphonophores.
Literature review
Different approaches of simulating nano-particles and possible adaptations for the proposed robotic platform are considered. Section 1
Simulating swarm behavior In literature, many robotic swarms are being used to simulate nanopar- ticle system behavior and targeted drug delivery. The advantage of using real robots rather than simulation is their ability to interact with the real world. In this scenario (nano-medicine), they can also serve the purpose of validating simplified sim- ulation results and complex dynamics of nanopar- ticles. In order to test their theory about bind- ing kinetics, Hauert et al., [14] divided a kilobot swarm[15] into two groups (one group represented the target and the other; nanoparticles). When initialised, the nanoparticle bots moved randomly until they found and boded with a cell bot. Al- though, the experiments actually resembled with the behavior of nanoparticles under the microscope, the swarm (of 1024 individuals) was not actually scaled to the population of nanoparticles (trillions).
Thus, a just comparison can’t be made to interpret the proposed hypothesis. This issue of scalability is synonymous to the problem of boundary cover- age where a swarm of agents is required to deliver some payload to a specific target [16]. The prob- lem arises when internal and external noises cause the same agents to behave differently when scaled. For instance, Brownian motion and chemical in- teractions have been known to affect the swarms differently at nano-scale when the population is increased [17]. This also makes it difficult to con- trol the swarms. In a survey paper, Chowdhury et al. [18] categorized control strategies for micro and nano-particles into two groups based on the particle size and the population 5. In both cases, it was observed that chaotic (unpredictable) behavior emerged even with a population of six particles at nano and micro scales. Furthermore, a huge re- search gap can be observed as shown in the Figure 3. This indicates that there is a need for better control strategies for the particles of nano and micro scales 6. It can also be noted that sensory noise (such
5
Using specialized substrates large number of microrobots can be actuated independently. However, their applica- tions are limited to workspaces that the substrate can reside.
Actuating microrobots with the motility of a living organ-
ismas iesaralynetwrigagpeprrodauceh tbouta tlheakinydetpuemndoern/ttacrognetrto)l aolfso
them and briefly discuss how they may help in addressing the current challenges.
Fig. 6 Maximum number of robots that can be controlled Figure 3: Maximum number of robots that can be
microorganisms is very difficult. There have been several
affects the robustness. This is specially the case
approaches ranging from optical to magnetic fields to con-
when the nanoparticles passively interact with the
trol the motility of the organisms which can be utilized to
tissue. According to a study of medical trial [19], control the microrobots. These hybrid approaches also suf-
more than 95 percent of passive nanoparticles end
fer from coupling where the heterogeneity among multiple7 upinspleen,liverandlungsduetotheEPReffect .
organisms are utilized to control them independently with a
An alternative approach can be to instead enforce
single global input.
morphological constraints and utilise the most ba- sic embodiment to overcome noise, scale and size
6.2 Future directions
effects. Although, literature on morphologically
Tocosnusmtmraairniztentahneof-uptaurteicdliersecftaioirnlsywliemhitaevde ilnoomkeeddicnitnoe,
ditffhiceuyltiheasvoef tshheocwurnresnotmapeprroeamchaesr.kWabelehapverodpiveirdteidesallin
the approaches into two categories: Independent control and
other fields. Mou Pal and co-workers harnessed
coupled control (Fig. 5). Coupled control utilizes the hetero- the shperical morphology of titania (TiO2), to pro-
geneity among the microrobots for simultaneous actuation
independently
duced uniform photonic crystals for dye senstised solar panels[20].
controlled independently[18]
..
1Independent control required no global input (self- organised) and opposite, coupled control exhibited parallel coordination with a global input.
2Note that Bacteria based approached are not included because they can’nt be used to realistically simulate the desired emergent behaviour [21].
3Rnhanced permeation and retention (EPR) effect refers to the increase in the retension time when the nanoparticles are coated with hydrophilic compunds
6
Timeline
April May June July August September
Weeks: 1 2 3 4 5 6 7 8 9 101112131415161718192021222324
50% complete
Overall
à IP C à à
Figure 4: Time-line legend
16% complete
30% complete 0% complete
0% complete
0% complete
0% complete
0% complete
30% complete
à Simulation, IP Robot design and parts arrival, C Aquarium
construction, à Assembly, Experiment 1, Experiment 2, Experiment 3, à Writeup,
Thesis
7
References
[1] R. H. Muller and C. M. Keck, “Challenges and solutions for the delivery of biotech drugs– a review of drug nanocrystal technology and lipid nanoparticles,” Journal of biotechnology, vol. 113, no. 1, pp. 151–170, 2004.
[2] O. Felfoul, M. Mohammadi, S. Taherkhani, D. De Lanauze, Y. Z. Xu, D. Loghin, S. Essa, S. Jancik, D. Houle, M. Lafleur, et al.,
“Magneto-aerotactic bacteria deliver drug- containing nanoliposomes to tumour hypoxic regions,” Nature Nanotechnology, 2016.
[3] L. E. van Vlerken, T. K. Vyas, and M. M. Amiji, “Poly (ethylene glycol)-modified nanocarriers for tumor-targeted and intracellu- lar delivery,” Pharmaceutical research, vol. 24, no. 8, pp. 1405–1414, 2007.
[4] P. Galvin, D. Thompson, K. B. Ryan, A. Mc- Carthy, A. C. Moore, C. S. Burke, M. Dyson, B. D. MacCraith, Y. K. Gun’ko, M. T. Byrne, et al., “Nanoparticle-based drug delivery: case studies for cancer and cardiovascular applica-
tions,” Cel lular and Molecular Life Sciences, vol. 69, no. 3, pp. 389–404, 2012.
[5] S. Hauert and S. N. Bhatia, “Mechanisms of cooperation in cancer nanomedicine: towards systems nanotechnology,” Trends in biotech- nology, vol. 32, no. 9, pp. 448–455, 2014.
[6] R. Pfeifer, M. Lungarella, and F. Iida, “Self- organization, embodiment, and biologically inspired robotics,” science, vol. 318, no. 5853, pp. 1088–1093, 2007.
[7] M. G. Hinchey, R. Sterritt, and C. Rouff, “Swarms and swarm intelligence,” Computer,
vol. 40, no. 4, 2007.
[8] Y. Katz, K. Tunstrøm, C. C. Ioannou, C. Huepe, and I. D. Couzin, “Inferring the structure and dynamics of interactions in schooling fish,” Proceedings of the Na- tional Academy of Sciences, vol. 108, no. 46, pp. 18720–18725, 2011.
[9] G. Beni, “From swarm intelligence to swarm robotics,” Swarm robotics, pp. 1–9, 2005.
[10] M. Marín, “Food, foraging, and timing of breeding of the black swift in california,” The
Wilson Bulletin, pp. 30–37, 1999.
[11] M. Salkoff, “Bees are swarming in the garden: a systematic synchronic study of productivity,” Language, pp. 288–346, 1983.
[12] R. M. Harshey, “Bees aren’t the only ones: swarming in gram-negative bacteria,” Molec- ular microbiology, vol. 13, no. 3, pp. 389–394, 1994.
[13] H. Aldrich, “The development of flagella in swarm cells of the myxomycete physarum flav- icomum,” Microbiology, vol. 50, no. 2, pp. 217– 222, 1968.
[14] S. Hauert, M. Rubenstein, S. B. A. Bagley, R. Nagpal, and S. Bhatia, “Insights from swarm-robotic experiments on the penetra- tion of targeted nanoparticles into tumors,”
[15] M. Rubenstein, C. Ahler, and R. Nagpal, “Kilobot: A low cost scalable robot system for collective behaviors,” in Robotics and Au- tomation (ICRA), 2012 IEEE International
Conference on, pp. 3293–3298, IEEE, 2012.
[16] K. Williams and J. Burdick, “Multi-robot boundary coverage with plan revision,” in Robotics and Automation, 2006. ICRA 2006. Proceedings 2006 IEEE International Confer- ence on, pp. 1716–1723, IEEE, 2006.
[17] S. P. Jang and S. U. Choi, “Role of brownian motion in the enhanced thermal conductivity of nanofluids,” Applied physics letters, vol. 84, no. 21, pp. 4316–4318, 2004.
[18] S. Chowdhury, W. Jing, and D. J. Cappel- leri, “Controlling multiple microrobots: recent progress and future challenges,” Journal of Micro-Bio Robotics, vol. 10, no. 1-4, pp. 1–11, 2015.
[19] Y. H. Bae and K. Park, “Targeted drug deliv- ery to tumors: myths, reality and possibility,” Journal of Controlled Release, vol. 153, no. 3, p. 198, 2011.
[20] M. Pal, J. Garcia Serrano, P. Santiago, and U. Pal, “Size-controlled synthesis of spherical tio2 nanoparticles: morphology, crystalliza- tion, and phase transition,” Journal of Physi- cal Chemistry C, vol. 111, no. 1, pp. 96–102, 2007.
[21] K. K. Jain, “Nanomedicine: application of nanobiotechnology in medical practice,” Med- ical principles and practice, vol. 17, no. 2, pp. 89–101, 2008.
8