Characterizing Protein Single Wall Carbon Nanotubes Dispersions and Rates of Cellular Uptake and Recovery

Single wall carbon nanotubes (SWCNTs) are attractive nanostructured materials for biological applications since they are chemically inert, mechanically strong, optically active, and have a high surface area-to-volume ratio. Natively, SWCNTs are highly hydrophobic and must be non-covalently exfoliated with biocompatible dispersing agents to preserve inherent SWCNT properties, reduce toxic bundles, and control SWCNT sub-cellular interactions. To generate biological SWCNT dispersions with the potential to control sub-cellular SWCNT-cell interactions, we dispersed SWCNTs with three different model proteins and directly compared quality, yield, and cellular uptake and sub-cellular localization. We investigated (i) bovine serum albumin (BSA) – the most prevalent blood serum protein whose interactions with SWCNTs have been previously characterized, (ii) lysozyme (LSZ) – a well-characterized, small molecular weight protein which has been investigated for use as a SWCNT dispersing agent, and (iii) γ-globulin – a set of immune proteins, which are “bioactive”.

            Direct comparisons – using NIR absorbance spectroscopy, Raman spectroscopy, and NIR fluorescence spectroscopy – of different protein dispersions demonstrated that these model proteins produced SWCNT dispersions with high yields (15-60%) while possessing near-infrared fluorescence of dynamic range up to ~50% that of the best synthetic (but cytotoxic) dispersing agents. Contrary to other reports, increasing the protein wt.% of the dispersion generally increases the quality of dispersion, although there is a threshold at ~0.5 wt.% above which quality only slightly increases with increasing yield. While increasing sonication time increases yield, it is associated with a decrease in NIR fluorescence dynamic range. This is surprising, as the Raman bundle peak is reduced with increasing sonication time; therefore, the reduction in SWCNT fluorescence may be due to sonication-induced SWCNT defects. Additionally, increased sonication time is associated with a decrease in protein secondary structure. To investigate how the different model protein dispersing agents interact with cells, we exposed NIH-3T3 cells (murine fibroblasts) to BSA-, LSZ-, and γ-globulin-dispersed SWCNTs. Interestingly, SWCNTs-LSZ and SWCNTs- γ-globulin are unstable in cell culture media. Quantification of SWCNT uptake reveals that SWCNTs-BSA are internalized at ~15 pg/cell while SWCNTs-LSZ are only internalized to ~1.5 pg/cell. This suggests that cellular efficacy cannot be determined solely by quality of the SWCNT dispersion.

            Next, we were interested in quantifying the time- and concentration-dependent rates of protein-dispersed SWCNT uptake, since a major obstacle to the development of controlled and efficacious SWCNTs for biomedical applications is the lack of quantification of cellular uptake and recovery rates. We determined the uptake and recovery of SWCNTs-BSA in NIH-3T3 cells, determining the mass of SWCNTs/cell by performing confocal Raman spectroscopy on cell lysates for concentrations ranging from 1– 100 µg/mL and times ranging from 5 s – 48 h. Interestingly, steady-state internalization of SWCNTs-BSA was reach in < 1 min, which is in agreement with modeling. As for the concentration rate, there was a threshold value between 1 and 30 µg/mL at which cell uptake machinery becomes saturated. To test the ability of cells to recover from SWCNT exposure, NIH-3T3 cells were sub-cultured and subsequently seeded onto fresh dishes with SWCNT-free media. As time increased, NIH-3T3 cells were able to expel SWCNTs-BSA, and after 80 h, cells were virtually free from intracellular SWCNTs. The ability of cells to recover holds promise for SWCNT-based biomedical applications, as healthy cells would be able to recover from systemic administration SWCNTs-BSA.

            As SWCNTs' excellent properties have led to tremendous interest in using SWCNTs for biological applications, we have directly compared model proteins dispersion capacity across a range of dispersion parameters. In general, proteins disperse SWCNTs while preserving SWCNT optical properties. However, different proteins lead to drastically different cellular internalizations. Also, we have shown that cells rapidly reach a steady-state internalization of SWCNTs-BSA within 1 min. However, cells recover at a different rate (31 ± 7 h), but are able, after ~80 h, to effectively expel SWCNTs. We believe these results will be greatly beneficial to the development of SWCNT-based nanobiological technologies. These technologies have potential applications in areas that include drug delivery, non-invasive high-contrast cell and tissue imaging, sub-cellular sensing, extra-cellular single biomolecule sensing, and controlling stem cell growth and differentiation.