This article is a contribution from Akhil Kallepalli*, Blake McCall, David B James, Sarah Junaid, James Halls, and Mark A. Richardson.
*School of Physics and Astronomy, University of Glasgow

A Novel Convergence of Biomedical Optics and Tissue Engineering Investigating Light Transport Through Synthetic Skin

1. Introduction
Human skin equivalents (HSE) are three-dimensional (3D) living models of human skin that replicate the biological properties and interactions within skin. They provide ideal testbeds for many applications such as wound healing assessment, drug assessment [1], [2], etc. and are now seeing application with patients in surgical theatres [3], [4]. For these lab-cultured tissues to be further applicable as mainstream procedures, their mechanical and optical properties must be comparable to in vivo skin. In our lab, we are interested in studying the nexus of tissue engineering and biomedical optics to observe and compare lab synthesis to the real-world analogy. In collaboration with Dr Junaid (Aston University), we focus on understanding the propagation of light through skin from both a simulation and experimental perspective. If you’re interested in more technical details, analysis and results from a recent study can be found in Kallepalli et al. 2020 [5].

2. Human Skin Equivalents
The human skin is a layered structure of cells at different stages of division. Amongst these layers, primarily, skin is divided into the superficial epidermis and underlying dermis layers. The epidermis is the first line of defense for the human body. The dermis layer serves the function of providing blood and the contained life-supporting oxygen and nutrients for the tissue. This structure begins with live cells at the basal layer (bottom layer of skin). The cells divide and go through their lifecycle as they move to the upper layers before eventually drying and falling off the body. This process is called terminal differentiation. The approach with human skin equivalents is to replicate this process by first generating the bottom layers of the skin before growing the rest of the model. For this purpose, the cells that compose the tissue in vivo are used and are known as primary cells. Fibroblasts that form the basement layers are cultured first, followed by live cells at different stages, as illustrated Figure 1. The keratinocytes form the epidermis, while the fibroblasts form the dermis layers, exactly as in the case of human skin. While the model is composed of live cells, it is important to note that there is a lack of melanin in the epidermis and haemoglobin in the dermis. Melanin is key for protection of skin against sunlight and is directly correlated to the skin color. Haemoglobin is present in the blood, bonds with oxygen and is responsible for transport of oxygen to tissue. Melanin and haemoglobin are the key absorbing entities in the tissue and are referred to as chromophores as they are responsible for the response to interaction with light. To confirm the successful culture of an HSE sample, histological analysis shows the stratified layers of the skin tissue (Figure 2). 

Figure 1 – The process of culturing stratified epidermis and dermis layers takes 4 weeks. Dermal fibroblasts and keratinocytes in this process.

3. Light interaction with skin
The effect of the human epidermis and dermis on the transmission and interaction with light depends on wavelength and resident chromophores (melanin, haemoglobin). Optically, the tissue is described in terms of its refractive indices and physical thickness. In combination, the optical properties and physical dimensions allow for a detailed simulation. 

Figure 2. Histological analysis of the HSE sample after optical experiments clearly shows the stratified epidermis and dermis layers.

All photo-biologic interactions are governed by the wavelength of the incident light, the intensity, and the time of exposure. Our spectral region of interest is limited on one end by the UV region (due to absorption by the epidermal melanosomes) and on the other end by the NIR radiation (>1020 nm) due to water absorption. Within this window that leaves us the visible and a small portion of the near-infrared wavelengths, most optical approaches to diagnostics are undertaken. Light encounters a mixture of reflection, absorption, scattering and transmission. The amount of light absorbed is governed by the concentration of the constituents of each tissue and the scattering albedo (ratio of absorption to scattering coefficients). 

In the epidermis, melanization (and the resulting skin color) is an important factor that affects the propagation of light into deeper layers. For normally incident light, between 4% and 7% of incident light is backscattered as regular reflectance (because the epidermal surface is not smooth and planar) for all skin types. The remaining 93% to 96% interacts with the tissue and results in scattering, transmission and/or absorption to varying degrees. Absorption by melanin is variable because it depends on the concentration, distribution, and thickness of the layers. In the near-infrared region, the backscatter from the epidermis is weak (compared to scattering) and the forward scattering mainly involves off-axis refraction and large-particle scattering [6]. In the turbid dermis layer, the primary chromophore is haemoglobin, and the dominant form of attenuation is scattering. The scattering coefficient is also inversely proportional to the wavelength, and longer wavelengths thus travel deeper into the dermis with less scattering. Greater transmittance of NIR wavelengths through the dermis and arterial blood allows medical diagnostics for measuring the pulse and heart rate and the amount of oxygen present in the blood [7].

4. Monte Carlo Methods
Monte Carlo simulations are the gold standard for modelling interactions between light and biological tissues and testing procedures in biomedical photonics [2], [8]. To achieve convergence to realistic results from stochastic methods, millions of interactions need to be accounted for. For interactions with ≥107 photons, ‘brute force’ MC simulations are impractical without variance reduction. This is achieved by importance sampling and ray splitting, which improves the efficiency and accuracy of the MC method. Importance sampling prioritizes the propagation of photons in a specific direction or onto a particular surface of significance. Ray splitting is used to enhance efficiency by splitting every interaction of a photon with an attenuating particle into four components: specular reflectance and transmittance, and scattered reflectance and transmittance. By also modelling absorption, the five components add up to the power of the incident photon. 

One of the best-known MC methods for biomedical optics in multi-layered tissues (MCML) [9] calculates the fraction of photon energy lost due to absorption using the absorption albedo (\left(\mu_a / \left(\mu_a + \mu_s\right)). Scattering events are quantified using polar and azimuthal angles calculated from the Henyey-Greenstein phase function. The photons are eventually eliminated when reflected or transmitted out of the tissue, or when their power drops below a predefined threshold. The photons are treated as random samples, whereas their absorption, scattering and transmission after interacting with chromophores and particles are physical processes. The optical properties of tissues are defined by the refractive index (\hat{\eta}), absorption coefficient (\mu_s), scattering coefficient (\mu_a) and anisotropy (g). The absorption coefficient and the refractive index define the material behavior, whereas the scattering coefficient and anisotropy influence the bulk scattering of light at a specific wavelength. The refractive index is defined in the our study as a complex index (\hat{\eta} = \eta + ik) where k = \lambda \mu_a / 4 \pi. Once the absorbed fraction of the photon’s energy is deducted, the remaining energy is attenuated based on the scattering distribution function (Henyey-Greenstein phase function, SDF), which is used in MC calculations.

    \[SDF = p\left(\theta\right) = \frac{1 - g^2}{4 \pi \left(1 + g^2 - 2 g \cos \theta\right)^{3/2}}\]

    \[g = \left<\cos \theta\right> \Rightarrow \theta = \left(g\right)\]

The dimensionless anisotropy (g) is the average cosine of the scattering angle . It represents the average scattering angle over numerous events [10]. Therefore, the variation in g dictates the scattering direction, with positive values indicating forward scattering, negative values indicating backscatter and zero for isotropy [11].

5. Analysis of light transport through human skin equivalents

The motivation of our research is to assess the comparability of human skin equivalents with in vivo skin. While considering the absence of the chromophores (analogous to tissue that has been removed from the body and prepared by washing with saline solution) and investigating previously published optical properties [12]–[16], we obtained the optical properties most relevant to our study with human skin equivalents. The principle behind this is highlighted by Mignon et al. [17] who cited a large range of optical properties in literature due to the multiple procedures of sample preparation and measurement. The analysis of the HSEs includes MC methods for comparing human skin (with and without chromophores) and experiments that measure the transmission in red and near-infrared wavelengths for comparison with the simulations. 

5.1. Simulated Optical Analysis
Simulated optical analysis considers the thickness of the skin layers along with adopted optical properties from the literature. The choice of the properties must be made carefully and specific to each study. The experiments show transmission to a higher degree due to the absence of the attenuating chromophores. 

5.2. Transmission Measures
When we ran our experiments, the twelve HSEs were illuminated with a laser beam, operating at red and near-infrared wavelengths (not at the same time). The light that interacts with the skin model and transmits through is measured by a power meter placed behind the sample (Figure 3). Each of the samples was then placed in a custom holder and mounted on a rail, directly in front of the laser source. The thickness of each sample was measured when the optical experiments were complete and were used to set up the simulation models. To compare and validate the thickness variations in the samples, the samples were prepared for haematoxylin and eosin (H&E) staining by washing with saline solution. H&E staining is a standard procedure when preparing samples for imaging under microscopes and gives images commonly associated with microscopy with pink/purple colors highlighting cells.

6. Results and Discussions
When the sample thickness is <0.3 mm, the transmission of NIR rays is unquestionably greater than red-colored light. Additionally, in the simulation and the experiments, we were able to ‘see’ the laser beam as opposed to a scattered speckle-like pattern. This led us to conclude that thin HSE samples (<0.3 mm) behave like filters, such that more light is absorbed than scattered, whereas the opposite applies for thicker samples due to the increased likelihood of interactions. When the thicker skin model was irradiated with red light, 3.2 mW (67%) of the power was absorbed, whereas the thinner models irradiated with NIR light absorbed only 12% of the incident power. The simulation of the thinnest sample forward scattered 55% of the red light and 60% of the NIR light. In contrast to the predictable effects on forward scatter, there was much greater variability in terms of backscatter. In thicker samples, 31% to 34% of the red light was backscattered compared to 38% to 52% of the NIR light. The thinner samples backscattered less than 30% of the incident light at both wavelengths, except for thin samples illuminated with NIR wavelengths. More light was absorbed by the thicker samples. Given optically identical layers (i.e., the same optical properties), forward scattering and the power of transmitted light increased for the thinner samples. The number of rays incident on the detector was also attenuated by the skin model at a given thickness and set of optical properties [13]. Human skin is characteristic in its interaction with light at different wavelengths. A key inference is the deeper propagation of photons into the tissue at longer wavelengths. Our study shows that the trend of transmission at red and near-infrared wavelengths is comparable to in vivo tissue. This finding makes the case for optical similarity of cultured and natural skin tissue and establishes a rapid approach for quality-checking the culture process. 

Figure 3 – The experimental setup measures the direct transmission of light through the HSE sample. The laser, sample and power meter are aligned, and the measured transmission is correlated with the simulation. 

Figure 4 – The comparison between simulation and experimental measures shows exponentially increasing differences as the sample becomes thinner (Pmeasured/Psim). The detected power of simulations with higher values of μa (-•-) is lower than a reported by Salomatina et al [27] (blue dashed line) until thickness of the sample falls below 0.5 mm. For these thinner samples, the detected power is almost identical

7. Conclusions
Our collaborative research lays the foundation for bringing tissue engineering and optical approaches closer together for diagnostics and standardizing the culture process. Further, we are keen on better understanding the division process of cells within the skin layers and studying the interaction of these cells in response to external stimuli. Wound healing is another avenue that will benefit from access to these HSE cultures as testbeds for understanding the intrinsic and biological processes governing the body’s response. Human skin equivalents are a state-of-the-art addition to any surgical or wound healing procedure. Should the cultures be optically and mechanically identical to human skin, their application potential is enormous. While challenges of speed of culture are now being dealt with by additive manufacturing approaches, the access to this cultured tissue could solve major surgical challenges such as those in maxillofacial reconstruction procedures.

8. Authors
Dr. Akhil Kallepalli is currently a research associate at the University of Glasgow, Dr. David James and Prof. Mark Richardson are at Cranfield University, Shrivenham, and Dr. Blake McCall and Dr. Sarah Junaid are at Aston University in Birmingham.


[1] M. Jannasch, F. Groeber, N. W. Brattig, C. Unger, H. Walles, and J. Hansmann, “Development and application of three-dimensional skin equivalents for the investigation of percutaneous worm invasion,” Exp. Parasitol., vol. 150, pp. 22–30, 2015, doi: 10.1016/j.exppara.2015.01.005.
[2] H. Mertsching, M. Weimer, S. Kersen, and H. Brunner, “Human skin equivalent as an alternative to animal testing.,” GMS Krankenhaushygiene Interdiszip., vol. 3, no. 1, 2008, [Online]. Available:
[3] W. H. Eaglstein, M. Iriondo, and K. Laszlo, “A Composite Skin Substitute (Graftskin) for Surgical Wounds,” Dermatologic Surg., vol. 21, no. 10, pp. 839–843, Oct. 1995, doi: 10.1111/j.1524-4725.1995.tb00709.x.
[4] Z. Zhang and B. B. Michniak-Kohn, “Tissue engineered human skin equivalents,” Pharmaceutics, vol. 4, pp. 26–41, 2012, doi: 10.3390/pharmaceutics4010026.
[5] A. Kallepalli, B. McCall, D. B. James, S. Junaid, J. Halls, and M. A. Richardson, “Optical investigation of threedimensional human skin equivalents: A pilot study,” J. Biophotonics, vol. 13, no. 1, Jan. 2020, doi: 10.1002/jbio.201960053.
[6] N. M. Maughan, J. W. Moody, and D. R. Miller, “Monte Carlo simulation of near-infrared light propagation through homogeneous mixed media,” J. Biomed. Opt., vol. 18, no. 10, p. 105007, 2013, doi: 10.1117/1.JBO.18.10.105007.
[7] A. Jubran, “Pulse oximetry,” Intensive Care Med., vol. 30, no. 11, pp. 2017–2020, 2004, doi: 10.1007/s00134-004-2399-x.
[8] T. Binzoni, T. S. Leung, A. H. Gandjbakhche, D. Rüfenacht, and D. T. Delpy, “Note on ‘The use of the Henyey-Greenstein phase function in Monte Carlo simulations in biomedical optics,’” Phys. Med. Biol., vol. 51, p. N313, 2006, doi: 10.1088/0031-9155/51/22/L01.
[9] L. Wang, S. L. Jacques, and L. Zheng, “MCML — Monte Carlo modeling of light transport in multi-layered tissues,” Comput. Methods Programs Biomed., vol. 47, no. 2, pp. 131–146, 1995, doi:
[10] I. J. Bigio and S. Fantini, “Overview of tissue optical properties,” in Quantitative Biomedical Optics (Theory, Methods, and Applications), 1st ed., Cambridge, United Kingdom: Cambridge University Press, 2016, pp. 19–59.
[11] M. J. C. van Gemert, S. L. Jacques, H. J. C. M. Sterenborg, and W. M. Star, “Skin optics,” IEEE Trans. Biomed. Eng., vol. 36, no. 12, pp. 1146–1154, 1989, doi: 10.1109/10.42108.
[12] S. L. Jacques, “Optical properties of biological tissues: A review,” Phys. Med. Biol., vol. 58, pp. R37–R61, 2013, doi: 10.1088/0031-9155/58/14/5007.
[13] R. Marchesini, C. Clemente, E. Pignoli, and M. Brambilla, “Optical properties of in vitro epidermis and their possible relationship with optical properties of in vivo skin,” J. Photochem. Photobiol. B Biol., vol. 16, pp. 127–140, 1992, doi: 10.1016/1011-1344(92)80004-F.
[14] I. V. Meglinski and S. J. Matcher, “Quantitative assessment of skin layers absorption and skin reflectance spectra simulation in the visible and near-infrared spectral regions,” Physiol. Meas., vol. 23, pp. 741–753, 2002, doi: 10.1088/0967-3334/23/4/312.
[15] E. Salomatina, B. Jiang, J. Novak, and A. N. Yaroslavsky, “Optical properties of normal and cancerous human skin in the visible and near-infrared spectral range,” J. Biomed. Opt., vol. 11, no. 6, p. 064026, 2006, doi: 10.1117/1.2398928.
[16] A. N. Bashkatov, E. A. Genina, V. I. Kochubey, and V. V. Tuchin, “Optical properties of human skin, subcutaneous and mucous tissues in the wavelength range from 400 to 2000 nm,” J. Phys. D. Appl. Phys., vol. 38, pp. 2543–2555, 2005, doi: 10.1088/0022-3727/38/15/004.
[17] C. Mignon, D. J. Tobin, M. Zeitouny, and N. E. Uzunbajakava, “Shedding light on the variability of optical skin properties: finding a path towards more accurate prediction of light propagation in human cutaneous compartments,” Biomed. Opt. Express, vol. 9, no. 2, p. 852, 2018, doi: 10.1364/boe.9.000852.