Engineered nanoparticles

Many of the structures within the human body, such as proteins, are measured in nanometers, making nanotechnology an alluring area of biomedical research. The growing ability of researchers to synthesize nanomaterials opens up the possibility of using synthetic molecules such as biosensors to probe cellular function. Biosensors based on optical approaches have undergone rapid growth; one example is the use of fluorescent tags for confocal microscopy and flow cytometry. Our group has recently been developing synthetic nanostructures as fluorescent biosensors to be used with fiber-optic probes for cancer diagnosis and treatment.

Conventional cancer treatment relies on the ability of chemotherapeutic drugs to preferentially kill rapidly dividing tumor cells. Most of these drugs, however, are generally toxic and induce serious side effects, providing ineffective therapy with little margin of error. The efficacy of cancer drugs is also often limited by their insolubility and instability, the low rate at which the tissue absorbs them, and a tumor's drug resistance. Doctors could lessen these problems to some degree if they could easily tell how the drug was distributed in a particular patient's tissue and how well it was killing tumor cells. Determining these facts, however, is difficult and requires complicated analytical techniques. If we could target cancer drugs directly to tumor cells and couple them with a biosensor that would analyze drug uptake and cell response, we would get much better results.

Polyamidoamine dendritic polymers offer an ideal platform of nanoscale dimensions that doctors could use to deliver drugs and biosensors to tumors. We felt that we could couple this platform to fluorescent dyes and other optically active substances, making it feasible to analyze drug uptake and efficacy by optical techniques.¹ The sensing molecule could be a fluorophore such as fluorescein.

To demonstrate this concept, we synthesized a dendritic polymer similar in size to a protein (molecular weight 25,000 amu and diameter 4 to 5 nm). We then covalently linked a tumor-targeting molecule (such as folic acid or HER2-antibody), a sensing fluorescent molecule, and a chemotherapeutic drug (methotrexate or taxol) to the polymer.

Figure 1. Fluorescent images show cancer cells incubated for 48 hours with either a non-targeted nanodevice (left) or a targeted nanodevice (right).

We purified and analyzed the drug-conjugated biosensors by multiple techniques to ensure they had proper structure and were free of contaminants. We tested the material to assure that the nanodevice therapeutics were stable in serum so they would not break down before reaching tumor cells and were biocompatible in animals, eliciting no host immune response or side toxicity. Research has shown the nanodevice to specifically target cancer cells in vitro and in vivo with binding characteristics consistent with those predicted by molecular modeling.² Confocal microscopic analysis provided evidence of cellular internalization of the device (see figure 1). A generation-5 dendrimer platform with fluorescein, folic acid, and methotrexate on its surface may be better at killing tumor cells in vitro than a free drug. We conducted further in vitro studies, using tumor cell pellets to replicate human tumors, to develop optical techniques for direct sensing of the multifunctional dendrimers.

Flow cytometry can detect and analyze fluorescent tags in tumor cells, but the method requires tissue processing and isolation of cells from the tumor prior to analysis. This is because tissue is a highly scattering medium that disrupts and strongly absorbs visible light. These characteristics render it difficult to detect fluorescence in real time by applying light sensors to the skin, particularly when analyzing any tissue depth beyond about a millimeter. We and other researchers have examined optical-fiber-based sensing technology as a minimally invasive approach to tissue analysis.^3-7 Our group decided to test optical fibers in conjunction with nanomolecular sensors for tumor analysis.

We opted for a new analysis technique for sensing in tumors, based on two-photon fluorescence (TPF) measurements through a thin optical fiber. The advantage of this technique over conventional one-photon fluorescence is better signal strength and discrimination using less light energy. We inject the nanomolecular biosensors into a vein and wait for them to migrate into the tumor cells in the test animal. Then we insert a fiber-optic probe into the tumor via a 27-gauge needle. This fiber senses the small excitation volume near the fiber tip as we pass it through the tumor mass, identifying and quantifying signals from the biosensor. The use of near-IR light lets us minimize photo damage to living cells, in contrast to using energetic UV photons. A large separation in wavelength between excitation and emission helps eliminate the detection background noise without sacrificing fluorescence signals from the biosensor. The technique lets us use a single laser source to excite a wide variety of fluorophores with different wavelengths of emission, which could allow simultaneous sensing of multiple signals.

We configured the biosensing system so that we could excite and collect TPF through the same optical fiber, which could allow real-time in vivo monitoring of biological systems. A titanium-doped sapphire laser provides 80-fs pulses at 830 nm with an 80-MHz repetition rate for two-photon excitation. The output tip of the fiber that carries the laser pulses is used to probe the biological samples. The same fiber collects the emitted TPF; a dichroic mirror separates the emitted signal from the excitation pulse, and a spectrometer resolves the wavelength. A photomultiplier tube detects the collected light after it travels through a short-pass filter.

We conducted initial experiments using a fluorescein solution as a standard and detected strong signals, which demonstrated our ability to excite and collect TPF through a single fiber. To find the optimal excitation- and emission-detection conditions, we tested different kinds of fibers, including single-mode, graded-index multimode, and step-index multimode. The fluorescence signal was higher from single-mode than multimode fibers, while the efficiency of graded-index multimode fiber was higher than that of step-index multimode fibers.

It is not immediately obvious why single-mode fibers led to larger signals, but compared to multimode fibers, they enable higher peak intensity at the fiber tip, which increases the two-photon excitation rate. The lower numerical aperture of single-mode fibers implies, however, that the fluorescence collection efficiency would be less than multimode fibers. To resolve this apparent paradox, we developed theoretical calculations to document the tradeoff between TPF excitation rate and collection efficiency. It turns out that the lower collection efficiency of the single-mode fiber is more than offset by the high peak power this fiber provides for efficient TPF excitation. This explained the experimental findings and validated the use of single-mode fibers in this application.

Figure 2. TPF data shows linear concentration dependence for standard nanosensor solutions (upper left). TPF (upper right) and flow cytometry data (lower left) from cultured KB cell pellets for non-targeted (green) and targeted (red) nanodevices shows effectiveness of the targeting mechanism. (UNIVERSITY OF MICHIGAN)

To determine the ability of this optical fiber to sense TPF in biological samples, we screened for the presence of a folic-acid-targeted nanodevice in cancer cell pellets that simulate human tumors. TPF measurement of standard nanosensor solutions using a single-mode fiber probe gave linear concentration dependence over a range of 10 nM to 100 µM (see figure 2). We then measured the TPF in cultured cell pellets treated with different concentrations of fluorescinated-targeted nanodevices. The targeted nanodevice shows significantly higher fluorescence than the non-targeted device. The observed binding was identical with data obtained by flow cytometry. The TPF data showed that at saturation, the level of nanodevices bound by the tumor cells was 2 pM per 10⁶ cells. These results demonstrate that the fiber-based TPF detection technique provides a promising method for in vivo biosensing of a variety of biological activities using fluorescence markers of different types and wavelengths. oe

References

1. P. Ferruti, R. Duncan, et al., in: Targeting of Drugs 6: Strategies for stealth therapeutic systems. G. Gregoriadis, B. McCormack (Eds.), Plenum Publishing Corp., NY, pp. 207-224 (1998).

2. A. Quintana, E. Raczka, et al., Pharmaceutical Research 19: 1310-1316 (2002).

3. A. Lago, A. Obeidat, et al., Opt. Lett. 20: 2054 (1995).

4. J. Sanghera, L. Shaw, et al., Proc. SPIE 3596: p. 178 (1999); A. Mulchandani, S. Pan, et al., Biotechnol. Prog. 15: 130 (1999).

5. S. Baker, Y. Zhao, et al., Anal. Chem. 71: 2071 (1999).

6. A. Abel, M. Weller, et al., Anal. Chem. 68: 2905 (1996).

7. J. Ye, M. Myaing, et al., Opt. Lett. 27: 1412 (2002).