3D nanostructural analysis of silk fibroin and recombinant spidroin 1 scaffolds by scanning probe nanotomography

Abstract

We use serial section scanning probe nanotomography to study the 3D structure, nanoporosity and pore interconnectivity of biocompatible scaffolds made of fibroin and recombinant spidroin 1 (rS1/9). Significant nanoporosity (24%) and pore percolation is detected for rS1/9 scaffolds. It is assumed to contribute to higher in vivo tissue regeneration efficiency of rS1/9 scaffolds.

---

A major challenge for tissue engineering is development of scaffolds that may substitute the native tissue matrix. Biocompatible and bio-degradable scaffolds provide a three-dimensional (3D) framework that can support cell attachment, proliferation and final differentiation into a target tissue, as well as 3D orientation of tissues.^1,2 The outstanding properties of silks such as biocompatibility, slow biodegradability and mechanical strength secure them as a basis for a variety of tissue engineering applications.^3,4 3D scaffolds produced from silk-based composites have shown promising results in vivo in various models of tissue regeneration.^5–7 The most comprehensive efforts so far have been directed toward the evaluation of specific properties as well as in vivo and in vitro behavior of natural silkworm and spider silks and their recombinant analogs, which were developed due to the limited natural sources of spider silks.^8–10

We have previously examined the physiological behavior of two scaffolds that were fabricated from different sources: the recombinant protein rS1/9 (an analog of spidroin 1, protein of dragline silk produced by Nephila clavipes spiders) expressed in yeast and fibroin from natural silkworm (Bombyx mori) silk using the same salt leaching technique.¹¹ The genetically engineered protein rS1/9 possesses all typical properties of full-length spidroins such as an ability for structural transition to the beta-sheet structure, a spontaneous formation of nanofibrils and microspheres in aqueous solutions, and formation of tough fibers after spinning.^12,13 Industrial scaling of such biotechnological manufacturing of recombinant analog of spidroin 1 from yeast may significantly decrease its production costs and make advantages of its unique properties cost-effective for wide use for tissue engineering.

Our study (including X-ray computed tomography and histological evaluation of retrieved specimens) indicated that the scaffolds from the spidroin-derivative recombinant protein rS1/9 demonstrate more effective bone tissue regeneration than the fibroin-based scaffolds. We believe that the observed differences between the behaviors of the scaffolds in vivo could be explained by the unique micro- and nanoporous structure that is formed during the fabrication rS1/9 scaffolds,^12,13 in contrast to the fibroin-based scaffolds.

The size, density and interconnectivity of micropores in scaffolds are critical for the efficacy of tissue integration; an interconnected structure is a prerequisite for homogeneous cell distribution and effective tissue intergrowth in vivo, as this type of structure facilitates active gas exchange, nutrient supply, and proper metabolic waste removal.^14,15 However 3D structure of such biocompatible porous scaffolds was not yet investigated at nanoscale. Study of size, density and interconnectivity of nanopores should provide important insights concerning their structural properties and help to further developing of the scaffolds with controlled biological and physiological properties.

There are several microscopy techniques that may provide information about 3D structure of porous biomaterials. X-Ray tomography is widely used nowadays for study of such materials on microscale.^16–19 But still resolution of this technique is not high enough to investigate needed nanoporosity properties of our scaffolds at scale of tens of nanometers. FIB/SEM technique is proved to be a versatile tool for nanoscale 3D reconstruction of various materials.^20–23 It is based on repetitive sputtering away material layers from the sample by the focused ion beam (FIB) technique and subsequent SEM imaging of the sample surface. Although, it may induce electron and ion radiation damage and unfavorable structural changes in studied soft biological and polymer samples.²³

An alternative approach for 3D nanostructural reconstruction is scanning probe nanotomography (SPNT) which is based on integration of scanning probe microscope (SPM) with an ultramicrotome²⁴ or a cryoultramicrotome²⁵ in one device. SPM is a surface characterization technique which is feasible for soft materials analysis without implying any considerable damage on the sample structures.²⁶ For 3D nanotomographical reconstruction of sample structures in the volume, consecutive SPM measurements are performed immediately at the fresh surface of an untreated block face after each ultramicrotome section. The depth resolution of 3D reconstruction thus is determined by the minimal thickness of repetitive ultramicrotome sections which may be as thin as 12 nm.^27,28 SPNT has been successfully applied for studying 3D structures of biological objects and materials, polymer blends and different nanocomposite and nanohybrid materials.^29,30 As soon as nanoporous materials are concerned SPNT analysis may provide the most direct assessment of nanopore structures in the volume by probing the pores on block face surface with SPM tip after each ultrathin section. Then 3D nanopore volume system may be reconstructed and visualized. However, this methodology has not been applied yet for comprehensive analysis of 3D nanostructure morphology of nanoporous materials.

In this work we use SPNT for investigation of 3D micro- and nanostructure of two porous scaffolds that were fabricated from the recombinant protein rS1/9 (analog of spidroin 1) expressed in yeast¹² and the natural silkworm fibroin respectively. For scaffold preparation the polymer proteins were dissolved in a solution of 10% lithium chloride in 90% formic acid (recombinant spidroins^31–33 and silk proteins^34–36 are known to be stable in low pH acidic environment). The porous scaffolds were prepared from biopolymer solutions (300 mg ml⁻¹) using the salt leaching technique. The proper amounts of dry NaCl particles of 200–400 μm in diameter were added to the polymer solutions up to 110 mg per 50 mcl and were mixed until they appeared to be homogeneously distributed. The shaped scaffold samples were dried at room temperature and then treated in 96% ethanol and immersed in distilled water to remove NaCl (see ESI† for more detailed description of scaffold preparation). Although both scaffolds were fabricated using the same technique and were shaped using the same molds, they displayed certain significant structural differences at micro- and nanoscale. It was shown that during the salt leaching procedure that is followed by a 96% ethanol treatment, silk-based materials adopt a beta-sheet structure. Further morphological evaluation revealed that macropore wall of rS1/9-based scaffolds spontaneously formed a microporous structure.^12,13 SEM evaluations demonstrated that the diameter of the macropores in both types of scaffolds lied within the same range of approximately 200–400 μm (Fig. 1). However, the scaffold pores in the fibroin and rS1/9 matrices showed differences in shape and in the depth of pore walls (Fig. 1a and b). It is not exactly clear why rS1/9-based and fibroin-based scaffolds display different microporous structure. These proteins have different amino acid sequences (the poly-A blocks and an alternation of hydrophobic and hydrophilic sections found in rS1/9 and the (GAGAGS)_n motif found the fibroin) and different molecular masses (94.3 kDa for rS1/9 and 370 kDa for fibroin). These differences can be the causes of different pore formation in scaffolds at the time of contact with the ethanol.

Fig. 1 SEM images of rS1/9 (a) and fibroin (b) scaffold macropores. Scale bar 100 μm.

Fig. 2 shows 2D AFM images of block face surfaces of epoxy-embedded spidroin and fibroin scaffolds after ultramicrotome sectioning. Analysis of 2D AFM images presented in Fig. 2 enables to estimate quantitatively the integral density and sizes of nanopores observed on the surface of sectioned scaffolds. Dimensions and other parameters of nanopores were calculated using GrainAnalysis software utility of NT-MDT Nova ImageAnalysis package. Although on both scaffolds we can detect pores with similar dimensions in range from 30 to 180 nm, and mean pore diameter of ∼50 nm, the integral density of nanopores detected on 2D AFM images and calculated volume porosity in rS1/9-based scaffold (46 μm⁻² and 24%) is dramatically higher than in the fibroin-based one (2.4 μm⁻² and 0.5% respectively). The spontaneously formed scaffold nanopore structure described in this study has not been reported so far.

The nanoporosity of rS1/9 scaffolds might be a specific feature of rS1/9 protein. Differences in micro- and nanoporosity between silk fibroin and rS1/9 may be caused by differences in molecular organization of these proteins. Recently in the study of Rouse-like rheological behavior of formic acid silk fibroin solution it was shown that formic acid exerted significant influence on the molecular weight of silk fibroin.³⁷ Differences in nanoporosity may also be caused by partial hydrolysis of silk fibroin during agitation in formic acid–LiCl solution and influence of hydrolysis products on the scaffold formation. We propose that further special investigations of influence of silk protein hydrolysis on its micro- and nanoposority are necessary.

Fig. 2 AFM images of block face surface of rS1/9 (a) and fibroin (b) scaffolds after ultramicrotome sectioning, 5.0 × 5.0 μm². Scale bar – 1 μm, scale bar in insert (a) – 200 nm, height variation – 30 nm.

One of the most important parameters is interconnectivity of nanopores in scaffolds. According to percolation theory interconnectivity of pores in the porous media starts to increase rapidly with porosity when the porosity is greater than a percolation threshold while almost no interconnectivity is observed for porosity below the threshold value.^38,39 For the percolation threshold of three-dimensional continuum media a universal volume ratio of 16% is usually considered.⁴⁰ Investigations and numerical simulations of various capillary porous materials report percolation threshold of volume porosity at 18–20% and high degree of interconnectivity for volume porosity of 24%.^41–43

Therefore we can expect that nanopores in rS1/9-based scaffold will have a significant degree of interconnectivity in three dimensions while no pore interconnectivity is detected in fibroin-based scaffolds. On 2D AFM images of the sectioned surface we can see that a noticeable part of the nanopores are connected with other one by nanochannels (see Fig. 2a, insert). However it is obvious that 3D analysis is necessary for adequate estimation of pore interconnectivity, especially considering that above-mentioned percolation threshold for three-dimensional systems is much smaller than the threshold for two-dimensional ones (∼44% volume ratio).³⁸ For investigation of 3D structure of scaffolds by SPNT technique we have performed serial ultramicrotome sectioning and consequent AFM measurements. For rS1/9-based scaffold sample 13 consequent AFM images of macropore wall cross-sections with 15.0 × 15.0 μm scan size and 1024 × 1024 pixel sizes each were acquired with 70 nm sectioning between images. Images were aligned in a stack and resulting 3D structure was visualized and analyzed with use of ImagePro AMS 6.0 software package with 3DConstructor option. Resulting images are shown in Fig. 3. For more comprehensive 3D visualization of nanopore system we use inclined sections through the reconstructed voxel volume (Fig. 3a and b) and pore volume image (Fig. 3c). Three-dimensional nanotomography reconstruction reveals interconnected pores and channels network in the bulk volume of rS1/9-based scaffold. Fig. 3b shows interconnection of individual nanopore channels through four consequent AFM images on the inclined section close-up. More detailed description of SPNT set-up and measurements is presented in ESI.†

Fig. 3 (a) SPNT 3D-reconstruction of rS1/9 scaffold macropore wall, overview, 15.0 × 15.0 × 1.0 μm³; (b) close-up of inclined section of 3D-reconstruction volume, scale bar – 500 nm, height variation – 30 nm; (c) SPNT 3D-reconstruction of pore volume in rS1/9 scaffold macropore wall, subvolume 4.68 × 4.0 × 1.0 μm³.

Both of studied epoxy-embedded scaffolds were sectioned well at room temperature and we detected no characteristic smearing or pronounced relief appearance on the block face surface which may be indicative that cryosectioning is needed. As series of consecutive AFM images were acquired the serial measurement was rather time consuming and took about 8 hours totally. Scan size, number of pixels, section thickness and number of sections were chosen as a compromise between total analysed volume, spatial resolution and measurement time. Together with some residual deformation of pores after sections it may lead to not exact matching of some nanopores in different images. However we are still able to make reasonable estimation of nanoporosity and nanopore percolation in the scaffold volume.

Processing of obtained SPNT 3D reconstruction data enable to visualize and characterize clusters of interconnected nanopores. Fig. 4 shows example of typical percolation pore cluster visualized by VolumeMeasurements utility of ImagePro AMS 6.0 software package. Dimensions of such clusters detected lay in range from 1 μm to 6 μm with the mean cluster size of 1.8 μm. Statistical analysis of 3D reconstruction of pores and pore clusters shows that porosity value is 23.8% which is consistent with the value obtained from 2D AFM images analysis. Volume fraction of pores interconnected in percolation clusters is 8.4% of the total volume or 35.3% of the total pore volume.

Fig. 4 Interconnected pore cluster in rS1/9 scaffold macropore wall, subvolume 5.86 × 2.91 × 1.0 μm³.

According to these results of 3D analysis we can expect significant nanoscale percolation and permeability in macropore walls which may positively effect to in vivo cell proliferation and corresponding tissue regeneration. Earlier we compared the ability of the silk fibroin and rS1/9 scaffolds to maintain an adhesion, distribution and proliferation of 3T3 fibroblasts. Both scaffolds demonstrated an equal ability to maintain a dynamic growth of the cells at 1, 4 and 14 days of culture on the fibroin and rS1/9 matrices.¹¹ But the vascularization and intergrowth of the connective tissue, which was penetrated with nerve fibers, at 8 weeks after subcutaneous implantation in Balb/c mice was more profound using the rS1/9 scaffolds. Based on the number of macrophages and multinuclear giant cells in the subcutaneous area and the number of osteoclasts in the bone, regeneration was determined to be more effective after the rS1/9 scaffolds were implanted.¹¹

We suggest that detected nanoporous inner structure of the rS1/9 scaffolds may contribute to a better micro-environment for the regenerating tissue, which makes the scaffolds derived from the recombinant rS1/9 protein more favorable candidates for in vivo applications. Although more investigations of nanoporosity correlation with biocompatibility properties and physiological behaviour of silk-based biomaterials are needed for exact estimation of its influence.

Presented nanotomography technique applied to analysis of micro- and nanoporous materials can provide resolution of tens of nanometers in all three dimensions which is comparable to resolution of FIB/SEM method but do not imply any electron or ion radiation damage in biological or polymer materials. Currently the technique is limited mainly by the range of materials analysed and the number of consequent images acquired in reasonable time but technical progress in development of cryoSPNT and fast-scanning SPM hardware and electronics tends to overcome these limitations significantly.²⁹ Further development of this technique is expected to provide important data on the structural organization of nanoporous systems in biomedical materials (tissue scaffolds and drug delivery microsystems), biological objects (tissues, cells, and microorganisms) and other nanocomposite materials like polymer electrolyte fuel cell membranes and catalyst layers.¹⁹ Therefore, it may become an indispensable tool for characterization of nanoporous materials and quality control of nanotechnological fabrication processes for a wide range of applications.

Acknowledgements

This work was supported by the Ministry of Higher Education and Science of the Russian Federation (Contract no. 14.604.21.0001, project unique identifier RFMEFI60414X0001). We also thank L. I. Davydova for carrying out some of the experiments.

Notes and references

1. L. Bacakova, E. Filova, F. Rypacek, V. Svorcik and V. Stary, Physiol. Res., 2004, 53(suppl. 1), 35.
2. J. Lee, M. J. Cuddihy and N. A. Kotov, Tissue Eng., Part B, 2008, 14, 61.
3. J. G. Hardy and T. R. Scheibel, Biochem. Soc. Trans., 2009, 37, 677.
4. G. H. Altman, F. Diaz, C. Jakuba, T. Calabro, R. L. Horan and J. Chen, et al., Biomaterials, 2003, 24, 401.
5. F. Sehnal and T. Sutherland, Prion, 2008, 2, 145.
6. A. Subramanian, U. M. Krishnan and S. Sethuraman, J. Biomed. Sci., 2009, 16, 108.
7. Z. Ge, Z. Jin and T. Cao, Biomed. Mater., 2008, 3, 022001.
8. C. K. Kuo, W. J. Li, R. L. Mauck and R. S. Tuan, Curr. Opin. Rheumatol., 2006, 18, 64.
9. J. A. Kluge, O. Rabotyagova, G. G. Leisk and D. L. Kaplan, Trends Biotechnol., 2008, 26, 244.
10. M. Heim, D. Keerl and T. Scheibel, Angew. Chem., Int. Ed. Engl., 2009, 48, 3584.
11. M. M. Moisenovich, O. Pustovalova, J. Shackelford, T. V. Vasiljeva, T. V. Druzhinina, Y. A. Kamenchuk, V. V. Guzeev, O. S. Sokolova, V. G. Bogush, V. G. Debabov, M. P. Kirpichnikov and I. I. Agapov, Biomaterials, 2012, 33(15), 3887.
12. V. G. Bogush, O. S. Sokolova, L. I. Davydova, D. V. Klinov, K. V. Sidoruk, N. G. Esipova, T. V. Neretina, I. A. Orchanskyi, V. Y. Makeev, V. G. Tumanyan, K. V. Shaitan, V. G. Debabov and M. P. Kirpichnikov, J. Neuroimmune Pharmacol., 2009, 4, 17.
13. M. M. Moisenovich, O. L. Pustovalova, A. Y. Arhipova, T. V. Vasiljeva, O. S. Sokolova, V. G. Bogush, V. G. Debabov, V. I. Sevastianov, M. P. Kirpichnikov and I. I. Agapov, J. Biomed. Mater. Res., Part A, 2011, 96(1), 125.
14. B. B. Mandal and S. C. Kundu, Biomaterials, 2009, 30, 2956.
15. J. Gao, P. M. Crapo and Y. Wang, Tissue Eng., 2006, 12, 917.
16. M. J. Moore, E. Jabbari, E. L. Ritman, L. Lu, B. L. Currier, A. J. Windebank and M. J. Yaszemski, J. Biomed. Mater. Res., Part A, 2004, 71(2), 258.
17. S. T. Hoa and D. W. Hutmacher, Biomaterials, 2006, 27, 1362.
18. A. Barbetta, R. Bedini, R. Pecci and M. Dentini, Ann. Ist. Super. Sanita, 2012, 48(1), 10–18.
19. L. Zielke, T. Hutzenlaub, D. R. Wheeler, I. Manke, T. Arlt, N. Paust, R. Zengerle and S. Thiele, Adv. Energy Mater., 2014, 4, 1301617.
20. A. Al-Abboodi, J. Fu, P. M. Doran and P. P. Y. Chan, Biotechnol. Bioeng., 2013, 110(1), 318.
21. L. Holzer, F. Indutnyi, P. H. Gasser, B. Muench and M. Wegmann, J. Microsc., 2004, 216(1), 84.
22. P. A. Midgley and R. E. Dunin-Borkowski, Nat. Mater., 2009, 8(4), 271.
23. A. Zankel, J. Wagner and P. Poelt, Micron, 2014, 62, 66.
24. A. E. Efimov, A. G. Tonevitsky, M. Dittrich and N. B. Matsko, J. Microsc., 2007, 226(3), 207.
25. A. E. Efimov, H. Gnaegi, R. Schaller, W. Grogger, F. Hofer and N. B. Matsko, Soft Matter, 2012, 8(38), 9756.
26. S. N. Magonov and D. H. Reneker, Annu. Rev. Mater. Sci., 1997, 27(1), 175.
27. A. Alekseev, D. Chen, E. E. Tkalya, M. G. Ghislandi, Y. Syurik, O. Ageev, J. Loos and G. de With, Adv. Funct. Mater., 2012, 22(6), 1311.
28. A. Alekseev, A. Efimov, K. Lu and J. Loos, Adv. Mater., 2009, 21(48), 4915.
29. A. Alekseev, A. Efimov, J. Loos, N. Matsko and J. Syurik, Eur. Polym. J., 2014, 52, 154.
30. N. B. Matsko and V. Mittal, Soft Matter, 2014, 10, 5478.
31. O. S. Sokolova, V. G. Bogush, L. I. Davydova, S. V. Polevova, S. A. Antonov, T. V. Neretina, D. V. Klinov, V. G. Debabov and M. P. Kirpichnikov, Mol. Biol., 2010, 44(1), 150.
32. L. A. Barr, S. R. Fahnestock and J. Yang, Mol. Breed., 2004, 13(4), 345.
33. J. Scheller, K.-H. Gührs, F. Grosse and U. Conrad, Nat. Biotechnol., 2001, 19, 573.
34. H. J. Cho, C. S. Ki, H. Oh, K. H. Lee and I. C. Um, Int. J. Biol. Macromol., 2012, 51(3), 336.
35. I. C. Um, H. Y. Kweon, Y. H. Park and S. Hudson, Int. J. Biol. Macromol., 2001, 29(2), 91.
36. I. C. Um, H. Y. Kweon, K. G. Lee and Y. H. Park, Int. J. Biol. Macromol., 2003, 33, 203.
37. Q. Wang, Q. Chen, Y. Yang and Z. Shao, Biomacromolecules, 2013, 14(1), 285.
38. A. Aharony and D. Stauffer, Introduction To Percolation Theory, Taylor & Francis, London, 2nd edn, 1994.
39. A. Hunt and R. Ewing, Percolation Theory for Flow in Porous Media, Lect. Notes Phys. 771, Springer, Berlin, Heidelberg, 2009.
40. H. Scher and R. Zallen, J. Chem. Phys., 1970, 53, 3759.
41. Y. N. Kryuchkov, Glass Ceram., 1997, 54(1–2), 58.
42. D. P. Bentz and E. J. Garboczi, Cem. Concr. Res., 1991, 21, 325.
43. J. Liu, K. Regenauer-Lieb, C. Hines, K. Liu, O. Gaede and A. Squelch, Geochem., Geophys., Geosyst., 2009, 10(5), 13.

---

Footnote

† Electronic supplementary information (ESI) available: Descriptions of preparation of scaffolds and scanning probe nanotomography and electron microscopy measurement set ups used. See DOI: 10.1039/c4ra08341e

---