Quantifying the mechanical properties of skin in vivo and ex vivo to optimise microneedle device design

The transdermal delivery of therapeutics is limited to only a few molecules due to the outermost layer of \nskin, the stratum corneum, which acts as a barrier against the ingress of substances into the body. \nMicroneedle arrays, which are commonly between 70μm and 900μm in length, have been developed as \na method of promoting drug and vaccine delivery by creating microperforations in the stratum corneum \nto increase transport into the skin. The design of microneedle devices has significantly developed over \nrecent years to allow for the delivery of numerous compounds into in vivo and ex vivo skin. Microneedle \ndevices are now beginning to be taken away from the laboratory and towards clinical use but to achieve \nthis it is desirable that all microneedles within the device penetrate skin in vivo to a sufficient depth. As \nmicroneedle devices have been extensively tested in cadaver tissue, a greater understanding of the \nmechanical properties of skin in vivo and ex vivo is required and to hypothesise whether animal models \nsuch as murine skin ex vivo serves as an appropriate model for human skin ex vivo. \nMeasurements were performed on human skin in vivo by applying small cylindrical and spherical \nindenters to the volar aspect of the forearm on 7 volunteers. The average Young’s Modulus of the skin \nwas 39.64kPa and 65.86kPa when applying the spherical and cylindrical indenters respectively. In a \nseries of tensile measurements performed at three load axis orientations using ex vivo samples from \nhuman and murine donors, it was found that the key variation was attributed to the deformation \nexperienced at initial low loads. This was shown to be significantly longer for human skin with an \naverage of 5.10mm, when compared with murine skin which had an average of 1.61mm (p<0.05). \nHistological examination showed that human skin was far thicker, with an increased volume of dermal \ntissue, compared with murine skin, and this anatomical variation may have been the main reason why \nhuman and murine skin exhibited different mechanical properties. \nFinite element models (FEMs) were established of skin indentation in vivo, which incorporated the \nepidermis, dermis and hypodermis, and of human and murine skin in tension. Appropriate boundary \nconditions and mesh densities were implemented and the geometries were taken from real life \nmeasurements where possible. The Ogden material model of hyperelasticity was chosen to represent \nthe skin layers for the FEM of skin indentation and an anisotropic material was used to describe human \nand murine skin in tension by adapting the Weiss et al model of transverse isotropy. \nInverse finite element analysis was then used to match the FEMs with the experimental measurements. \nThe multilayered FEM of skin was correlated against the in vivo indentation tests where model and \nexperimental fit gave average root mean squared errors (R2 \nave) of between 0.00103 and 0.0488 for the 7 \nvolunteers. The optimal material parameters showed correlations with experimental measurements, \nwhere volunteers 1, 6 and 7 were shown to have the stiffest skin through Young’s Modulus calculations, \nwhich was reflected in the increased nonlinearity of the parameters extracted for the hypodermal layer. \nA stronger agreement between model and experiment for the anisotropic model of human and murine \nskin in tension was shown where the R2 \nave was between 0.0038 and 0.0163. Again, model and \nexperimental observations were shown to correlate where there was a significant difference (p<0.05) \nbetween 6 of the 14 average material parameters (C2, C3,1, λ1, C3,2, C3,3, λ3) when comparing human to \nmurine skin. \nThe multilayered FEM of human skin in vivo was further validated by modelling the application of a \nsingle microneedle to skin, prior to penetration. The model was then correlated against in vivo \nmeasurements performed on one of the volunteers and it was found that the model provided a good \napproximation for the experimental measurements. Using the multilayered FEM of human skin \nindentation, it was possible to model the deflection of the skin during the application of a pressure load \ncomparable to microneedle array application. This allowed for the development of several curved \nmicroneedle arrays which aimed to distribute the load over all microneedles to potentially create uniform \nskin penetration by all those within the array. The microneedles were manufactured simply and quickly \nusing wire cutting technologies from stainless steel and tested in human skin in vivo and in ex vivo \nsamples of human and murine skin, where methylene blue was applied to identify any microchannels \ncreated by the microneedles. Preliminary measurements taken from murine skin ex vivo were \ndiscounted as microchannel staining was not possible. Analyses performed on human skin ex vivo \nshowed penetration at high loads (4-5N) for all four microneedle array designs and the microneedle \narray with the smallest curvature (0.95mm) had the most consistent puncture for all microneedles, \nhowever puncture in vivo was difficult to characterise using approach developed. Therefore further work \nis required to assess more volunteers and donors. \nThis study has highlighted the great differences in the mechanical properties of human and murine skin, \nsuggesting that murine skin is not an appropriate model to assess microneedle puncture. It has also \nshown that the underlying tissues and hypodermis play a pivotal role in microneedle insertion mechanics.