Review of Long-Wavelength Optical and NIR Imaging Materials: Contrast Agents, Fluorophores and Multifunctional Nano Carriers

Abstract

The importance of long wavelength and near infra-red (NIR) imaging has dramatically increased due to the desire to perform whole animal and deep tissue imaging. The adoption of NIR imaging is also growing rapidly due to the availability of targeted biological agents for diagnosis and basic medical research that can be imaged in vivo. The wavelength range of 650–1450 nm falls in the region of the spectrum with the lowest absorption in tissue and therefore enables the deepest tissue penetration. This is the wavelength range we focus on with this review. To operate effectively the imaging agents must both be excited and must emit in this long-wavelength window. We review the agents used both for imaging by absorption, scattering, and excitation (such as fluorescence). Imaging agents comprise both aqueous soluble and insoluble species, both organic and inorganic, and unimolecular and supramolecular constructs. The interest in multi-modal imaging, which involves delivery of actives, targeting, and imaging, requires nanocarriers or supramolecular assemblies. Nanoparticles for diagnostics also have advantages in increasing circulation time and increased imaging brightness relative to single molecule imaging agents. This has led to rapid advances in nanocarriers for long-wavelength, NIR imaging.

I. Introduction

Broadly, the field of biomedical imaging can be divided into categories based upon the electromagnetic spectrum as shown in Fig. 1: magnetic resonance, optical/near infra-red (NIR), and ionizing radiation (X-rays and gamma rays). Imaging based on ionizing radiation generally refers to the detection of high frequency emissions from radioactive elements such as the gamma ray emitters ¹¹¹In or ^99m Tc or the passage of X-rays through the body. The main technologies involved are positron emission tomography (PET)^1–6, single-photon emission computed tomography (SPECT)^7–9, and X-ray computed tomography (CT).^{1, 2, 7, 8, 10, 11} Magnetic resonance imaging (MRI) tends to operate on the other end of the spectrum in the MHz frequency range, relying upon contrast agents such as gadolinium or superparamagnetic iron oxide to modify the relaxivity of water molecules to provide soft tissue contrast.^{4–6, 12, 13}

Fig. 1.

The electromagnetic spectrum (adapted from Richards (2001)).³¹

The focus of this review is materials for optical and near infrared (NIR) imaging for diagnostic and therapeutic applications in biology and medicine. The important emerging area of whole animal and deep tissue imaging has made long-wavelength and NIR imaging of special significance. While optical and NIR wavelengths range from 400 nm up to 2500 nm, in vivo biomedical imaging requires consideration of the so-called 'imaging window', shown in Fig. 2. The figure shows optical absorbance as a function of wavelength and the attenuation arising from various components in a representative tissue sample. The absorbance units are cm⁻¹ and, therefore, the inverse of this absorbance represents a characteristic penetration depth. Absorbance by water, melanin, proteins, and hemoglobin (Hb) are high between 200–650 nm, covering essentially the whole visible range. In addition to absorbing light, tissues can also reflect, refract, and scatter incident photons.^14–16 While this does not create a problem for surface imaging of thin sections or single cell layers, it does point out the problem of imaging through tissue. Within the 200–650 nm range, autofluorescence from tissue also confounds fluorescence measurements^{17, 18} by creating high background fluorescence. It is with this realization that much attention has been focused in recent years on the development of absorption imaging agents and fluorophores with absorbance or excitation/emission maxima falling in the region of minimal tissue absorbance/autofluorescence between 650–1450 nm, an ‘imaging window’. Tissues have minimal absorbance in this wavelength range allowing for deep penetration of light. This enables whole animal imaging with high sensitivity in core organs in real time without the need for dissection. Therefore, we focus on dyes, imaging agents, and imaging nanocarriers in the wavelength range 650–1450 nm. While several excellent reviews cover various aspects of imaging,^19–30 there has not been a review that covers long wavelength and NIR imaging materials adequately.

Fig. 2.

Absorbance of various tissue and blood components from 200 nm to 10 µm. The optical imaging window ranging from 650 – 1450 nm represents the range where tissue penetration is greatest. This wavelength range is the focus of this review. Adapted from Boulnois (1986)³² and Susi (1971).³³

While the intrinsic properties of a particular absorber or fluorophore (excitation/emission wavelengths, quantum yield, absorption coefficient, photostability, chemical stability in different environments) play a large role in determining the effectiveness of an imaging agent, the method of delivery can have profound effects as well. Dyes can be delivered as soluble molecules in the bloodstream^{34, 35} or encapsulated within nanoparticles that may be polymeric or inorganic in nature.^36–40 The desired nanocarrier size depends strongly on the type of imaging and target organ. Generally, any size is acceptable for GI tract imaging (oral administration) while aerosol targeting to the lungs requires an upper size limit on the order of tens of microns depending on particle density for capture in the deep lung (see Fig. 4).^{41, 42} If the carrier is being administered via intravenous injection, particles greater than 1 µm can cause blockage at the injection site. For broad biodistribution, particles less than ~30 nm in size can penetrate most cells but are cleared relatively rapidly through renal filtration. Nanocarriers from 50–400 nm in size with the appropriate ‘stealth’ surface chemistry provide enhanced circulation time and passive accumulation in tumors due to the enhanced permeation and retention effect (EPR).⁴³ This non-specific targeting arises from the leaky vasculature that surrounds solid tumors with poor lymphatic drainage, with the result that nanocarriers enter the tumor site but are not rapidly cleared. Dyes can also be used to label antibodies for combined targeting and imaging capabilities.⁴⁴ The absorber or fluorophore may itself form nanoscale structures such as clusters, spheres, or rods which can then be used directly or further stabilized by polymers or inorganic agents.^45–47

Fig. 4.

Lung clearance of monodisperse polystyrene particles in rats (3 µm | 9 µm | 15 µm). Radioactively tagged polystyrene particles, intratracheally administered, were monitored for clearance from the lung. Adapted from Oberdörster (1989) and Snipes (1981).^{41, 42}

In this review, we highlight the development of imaging agents consisting either of pure absorbers, scatterers, or fluorophores that both absorb and re-emit light. Absorbers include traditional highly absorbing materials that attenuate optical radiation. Optical attenuation can also occur by scattering from agents with strong dielectric mismatch with tissue. Detection of scattered light for imaging requires more instrumentally and algorithmically challenging approaches. These rely on more complex interactions with optical radiation including phase, transit time, and polarization information. The reader is directed to excellent reviews of these imaging techniques.^{14–16, 48, 49} The agents that produce the dielectric mismatches required for these imaging techniques are similar to those required for imaging by attenuation, so we do not treat them as distinct imaging materials.

We review traditional organic absorbers and fluorophores, hybrids such as Zn²⁺ -multiporphyrin complexes,³⁹ ICG-doped silica-shell Au nanoparticles,³⁸ and novel inorganic materials such as quantum dots and upconverting phosphors.

While research into novel fluorophores and imaging agents has greatly expanded, the reality is that equally important are the delivery vehicles for these agents. Novel and efficient delivery agents are needed to provide stability, longer circulation time, as well as therapeutic and targeting capabilities, all in one multifunctional package. As such we will conclude by examining the state of the art in the new field of ‘theranostics’ which combines therapeutics and diagnostics.⁵⁰

II. Near-IR Absorption Imaging Agents

Imaging agents based on absorbance, or optical attenuation, require transmission of light through the entire sample. Even in the “imaging window” wavelengths, the absorption coefficients in Fig. 2 show that this can be challenging. Both organic and inorganic compounds can be used to provide contrast through photon/electromagnetic absorbance. These contrast agents have the advantage of not suffering from quenching effects that are seen with fluorescent agents. Extinction coefficients are relatively insensitive to chemical environments since the attenuation is due to internal electromagnetic structure and not transitions between states.

Gold (Au) nanoparticles exhibit a size- and morphology-dependent surface plasmon resonance absorbance band that is centered at 520 nm for 5 nm particles and shifts into the NIR region for larger spherical clusters or non-spherical morphologies such as nanorods.⁵² As seen in Fig. 5, the aspect ratio (AR) of gold nanorods (ellipsoids) strongly affects the plasmon resonance absorption band. It is predicted that an AR of ~3 is required to have absorption at 700 nm.⁵¹ The applicability of Au species for imaging in most cases requires obtaining adequate size in at least one dimension to shift the plasmon frequency to NIR wavelengths. But a problem with application of these larger colloids in vivo is ultimate clearance or toxicity associated with mechanical blockage. The Johnston group has addressed this problem by creating pH-dependent, polymer-protected, reversible gold nanoclusters from 5 nm Au colloids. These clusters can be tuned to have well defined diameters between 30–100 nm. This size control is primarily achieved through controlling the gold:polymer ratio, where the polymer is a PLA(2k)-b-PEG(10k)-b-PLA(2k) triblock copolymer that adsorbs weakly onto the gold surface and modifies the energetics of clustering. These clusters have high gold loadings with only 3% polymer by mass of the cluster. Once internalized into endosomes or other structures with low pH, the PLA degrades over time releasing the clusters back to individual Au nanoparticles thus allowing for fast renal clearance.^{46, 47}

Fig. 5.

Predicted absorption spectra for gold ellipsoids with varying aspect ratios (ARs). As seen, increasing the AR monotonically shifts the plasmon resonance band into the NIR, which is ideal for deep tissue imaging. Reproduced from Link, et al.⁵¹

Polypyrroles (PPy) are highly conducting materials due to their highly delocalized electrons which cause them to exhibit strong, broad absorption bands from 800 to ~2500 nm, thus falling into the ‘imaging window’.^56–58 The doped PPys have stronger absorption than the undoped neutral forms. The reader is directed to appropriate reviews for a discussion of the mechanisms responsible for these differences.^{59, 60} They display good chemical and thermal stability. While they are most commonly used in organic electronics applications,^{26, 57, 58} they are structurally and chemically related to the phthalocyanines and porphyrins which have already found widespread application as photosensitizers for photodynamic therapy (see Fig. 6).

Fig. 6.

Adapted chemical structures for A - pyrrole⁵³, B - polypyrrole⁵⁴, C - porphyrin⁵⁵, and D - phthalocyanine (adapted from Sigma-Aldrich Co. LLC, St. Louis MO).

Bjorklund, et al.⁶¹ published one of the first instances of colloidally stable PPy nanoparticles in the size range 100–200 nm. These particles were composed of pyrrole units polymerized in the presence of methylcellulose.⁶¹ Concurrently, Armes and coworkers produced 100–150 nm monodisperse PPy nanoparticles stabilized with poly(vinyl acetate) (PVA). The stability of these particles depended strongly on the amount of adsorbed PVA.⁶² Armes also created larger core-shell PPy particles using PPy coated poly(styrene) (PS) and poly(methyl methacrylate) (PMMA) latices.^{63, 64} In the former case, commercially available, charge-stabilized PS latices were coated with poly(vinylpyrrolidone) (PVP) to provide steric rather than charge stabilization. Then, smooth layers of PPy were deposited onto these 1.6–1.8 µm beads, adding less than 20 nm in thickness to the original particles. Colloidal stability was demonstrated for PPy loadings up to 9.9 wt% of the latex mass, beyond which flocculation and instability was observed. Ormond-Prout, et al. deposited PPy onto PMMA latices from 1.2 µm in diameter up to 30 µm with loadings from 3% to 21% by mass. As before, the PPy layer added no more than 20 nm to the particle diameter but had a distinctly globular surface morphology due to the less hydrophobic nature of PMMA compared to PS.⁶⁴

A recent study has shown the feasibility of targeting carboxylated PPy latices and PPy-silica composites with antibodies. These systems were shown to have similar immunoactivity thus opening the way for future targeting studies.⁴⁵ In a similar vein, Bousalem, et al. formulated 1.1 µm PS latices coated with PPy and conjugated to human serum albumin (HSA). Their studies indicated that the surface immobilized HSA retained its immunoactivity against anti-HSA.⁶⁵ Though PPy would seem to make a good biomedical imaging agent,⁶⁶ we are not aware of any reports on optical/NIR imaging of PPy-containing imaging agents in an in vivo setting. Though the toxicity of PPy is strongly dependent on the method of polymerization and specific biological setting, many reports have indicated that PPy has minimal negative effects on biological tissues.⁶⁷ Most of the reports of PPy in animal models focus on their electrical properties rather than the strong optical/NIR absorbance.⁶⁸

Taking a different approach, Jang, et al. developed novel water soluble, photofunctional, charged dendrimers with porphyrin or phthalocyanine cores that, when complexed with poly(L-lysine)-b-poly(ethylene glycol) (PLL-b-PEG) or poly(aspartate)-b-poly(ethylene glycol), form charge-neutral micelles on the order of 50 nm in size. These entities were shown to be stable under physiological conditions and are expected to be long-circulating in vivo due to the PEG surface.⁶⁹ The phthalocyanine core dendrimer (λ_abs = 685 nm) proved to have better spectral characteristics for biomedical applications than the porphyrin core (λ_abs = 560 nm).⁷⁰ It is unlikely that porphyrins/phthalocyanines would be used solely for imaging purposes as they are potent producers of reactive oxygen species (ROS) which are cytotoxic to cells. Other approaches involve the synthesis of increasingly large porphyrin rings as Tanaka, et al. have demonstrated (Fig. 7). From the basic porphyrin up to octadecaphyrin, the absorption band shifts from 411 – 953 nm, well into the NIR window.⁷¹ Srinivasan, et al. recently introduced bis-metal complexed hexaphyrins with ‘confused’ pyrrolic units which allow for easy complexation of different types of metals. Absorption bands from 325 – 755 nm were observed for the uncomplexed molecule while the bis-Ni compound had absorption bands from 360 – 1210 nm, well into the NIR range.⁷²

Fig. 7.

Increasingly large porphyrin rings have absorption maxima that shift into the NIR (411 – 953 nm). A – porphyrin, B – dodecaphyrin, C – octadecaphyrin (Ar = C₆F₅). Reproduced from Tanaka, et al.⁷¹

III. Near-IR Fluorophores

Fluorescent imaging is the most widely used imaging modality for biomedical research and diagnostics. Fluorophores can be either organic or inorganic and each has its advantages and drawbacks. As fluorescent imagers by definition involve both excitation and emission events (see Fig. 3), it is desirable for in vivo applications that the excitation and emission peaks fall into the ‘imaging window’ (650 – 1450 nm) where absorption by tissues and blood is at a minimum. Otherwise, even signals from high quantum yield fluorophores will be severely attenuated either from inadequate excitation or absorption of emission.

Fig. 3.

Jablonski diagrams showing the main energy pathways for different modes of fluorescence (energy diagrams not drawn to scale). A – single photon Stokes process (i) photon absorption gives excited state, (ii) internal conversion to S₁, first singlet excited state, (iii) fluorescence (emission of photon), (iv) nonradiative decay, (v) intersystem crossing to T₁ (‘forbidden’ triplet excited state), (vi) phosphorescence, and (vii) nonradiative decay.²⁴ B – two-photon upconversion (anti-Stokes process). Excited-state absorption (ESA) proceeds via sequential absorption of two photons to give the excited state and a subsequent emission event. In energy transfer upconversion (ETU), an ion directly absorbs one photon while a neighboring ion absorbs another and transfers the energy to the first ion, resulting in upconversion and emission.²² Reproduced from Lavis and Raines²⁴ and Haase and Schafer.²²

Inorganic fluorophores

Inorganic fluorescent agents fall into the broad categories of quantum dots (QDs) or upconverting phosphors (UCPs)

Quantum dots usually have a core-shell architecture comprising semiconductor materials such as CdSe, CdTe, PbS, or some alloy of those materials, with the composition strongly affecting the excitation and emission characteristics. As most of these materials have well-established toxicities, a shell of ZnS is often added to passivate the surface and improve the quantum yield and stability.⁷³ Water solubility and additional stability is conferred by coating the QD with amphiphilic copolymers, block copolymers, or other water soluble ligands.⁷⁴ QDs usually range in size from 2 to 30 nm (depending on the core size and shell characteristics), with larger cores emitting at longer wavelengths.^{73, 74} Theoretical modeling has shown two possible emission bands for in vivo imaging: 700–900 nm and 1200–1600 nm. While most QDs have excitation and emission spectra at shorter wavelengths, Kim, et al. and Park, et al. reported the synthesis and in vivo testing of QDs with ‘imaging window’ wavelengths.^{75, 76} QDs have broad excitation bands with narrow emission bands, allowing for multiple-emission agents to be detected at different wavelengths simultaneously with only one excitation source. They are generally bright and photostable with quantum yields from 20% to 60% and do not suffer from the quenching observed for organic fluorophores.⁷⁷

Gao and coworkers have demonstrated the feasibility of targeted bioconjugated QD probes for fluorescence diagnostics (see Fig. 8). Their CdSe core, ZnS shell, triblock copolymer protected nanoparticles are capable of 535–630 nm emission profiles and are about 30 nm in diameter. Significantly, they also demonstrated the targeting capabilities of these fluorescent probes by attaching antibodies for the prostate-specific membrane antigen (PSMA). In this way, tumors were identified by active targeting mechanisms. Due to the non-optical window wavelengths, fluorescent light detection within the liver and spleen was very limited.⁷⁷

Fig. 8.

In vivo fluorescent images of targeted QDs in mice. A – Targeted QDs preferentially accumulate in tumors. B – With a single excitation source, QDs tuned to different emission wavelengths can be detected simultaneously. Reproduced from Gao, et al.⁷⁷

Kim, et al.⁷⁵ and Park, et al.⁷⁶ reported QDs with longer wavelength excitation and emission. In the former case, type II QDs were synthesized with an oligomeric phosphine coating to confer water solubility. These QDs had a hydrodynamic diameter of 16 nm with 850 nm emission and a broad excitation band from 800 nm down to 500 nm. The QDs were assessed as real-time surgical aids and were found to accumulate preferentially in sentinel lymph nodes due to their size in both small and large animals (mice and pigs). This lymph node accumulation is expected to be very useful in the diagnosis and treatment of breast and colon cancers.⁷⁵ Park, et al. synthesized 745 nm emitting QDs comprising CuInSe in the core with ZnS as the shell. The average QD diameter was 4 nm with up to 60% quantum yield after ZnS passivation. Acyl chain lipids with PEG units were used as the water transfer agent which increased the hydrodynamic size to 15 nm. In water, the QDs were stable for several days at room temperature. Mouse tail-vein injections showed relatively uniform biodistribution with a clearance half-life of 286 mins.⁷⁶ No active targeting approaches were attempted in these studies.

Upconverting phosphors (UCPs) are a fundamentally different class of inorganic fluorophores that have recently improved to the point where they are now strong alternatives to quantum dots and organic fluorophores.²² Upconversion is a nonlinear optical process where sequentially absorbed photons are emitted at a shorter wavelength (and thus a higher frequency/energy) thus making it an anti-Stokes process (Fig. 3). The most efficient upconverting phosphors are hexagonal phase (β) crystals composed of rare earth dopants (lanthanides such as Yb³⁺ and Er³⁺) that provide the fluorescence while the solid state host matrix (NaYF₄) optimally positions these ions for energy transfer (Fig. 9). As these fluorescent centers are permanently separated in a solid state matrix, UCPs do not suffer from the same concentration dependent quenching effects as organic fluorophores and do not photobleach.²² Quantum yields are not defined for upconverting phosphors due to non-linear emission energy scaling with excitation energy.⁷⁸ The quadratic dependence of emission means that high illumination power is required for bright emission. Fortunately, the telecommunications industry, with solid state lasers in the wavelength range of 980 nm used for fiber optic communications has made high power very inexpensive in this wavelength range. The excitation of the crystal lattice is quenched by surface defects, therefore, emission intensity scales inversely with crystal size.⁷⁸ While 100 nm particles are readily imaged, imaging particles below 30 nm has proven challenging.⁷⁸

Fig. 9.

TEM image of oleic acid-trioctylphosphine stabilized NaYF₄:Yb³⁺,Er³⁺ upconverting nanophosphors. Reproduced from Budijono, et al. (2010).³⁷

Budijono, et al.³⁷ and Shan, et al.⁷⁹ have demonstrated the production of stable, multifunctional, ~200 nm diameter polymer-protected UCP nanoparticles without a photosensitizer (for diagnostic imaging) as well as with a photosensitizer (for photodyanamic therapy). ^{37, 79} These particles have a 980/540,660 nm ex/em profile. The photosensitizer (tetraphenylporphyrin) absorbance profile strongly overlaps with the UCP emission profile at 540 nm to enable reactive oxygen generation by the porphyrins for photodynamic therapy.

Nam, et al. have reported the feasibility of using UCPs as imaging agents in biological systems. Their work involved the synthesis of 40 nm PEG-phospholipid coated UCPs that were taken up by HeLa cells in vitro. At 980 nm laser excitation, the cells did not experience adverse effects and the UCPs were clearly visible at a high concentration within the cells. Cytotoxic effects from the UCPs were negligible and photobleaching was not observed even after 6 hours of continuous tracking.⁸⁰

Coating for the UCP crystal surface is required for biocompatibility and several approaches have been advanced. There are two promising routes to biocompatible surfaces for in vivo applications. The first by Hildebrand and Vinegoni involves polyacrylic acid coating followed by amine conjugation of PEG onto the polyacrylic acid coating to make the coated UCP bio-inert. ^{81, 82} The second route by the Prud’homme group involves block-copolymer-directed assembly with a PEG diblock copolymer to provide a dense PEG surface layer.^{37, 79} The directed assembly process is described below in Section IV.

Organic fluorophores

Most organic fluorescent agents with NIR excitation/emission profiles fit into one of four categories: cyanines, phthalocyanines/porphyrins/pyrroles (see section IV),³⁹ squaraines, or BODIPYs.²⁶ Other categories of dyes with non-NIR wavelengths have been well covered by Lavis and Raines.²⁴ A wide variety of absorbers and fluorophores have been cataloged and summarized in Table 1.

Table 1.

Absorption or excitation/emission characteristics and chemical/spectroscopic properties of imaging agents where λ_abs is/are the absorption maximum/maxima and ε is the molar extinction coefficient. Average extinction coefficients are given for some of the major classes, in the case the extinction coefficient for a specific compound is not available. Where multiple excitation and emission wavelengths are reported, the values of excitation maxima and emission maxima are given in order.

Absorber	λabs (nm)	ε (M−1 cm−1)	Chemical Properties	Reference
Gold nanoparticles/nanorods/clusters	Broad 700– 900 (size- and aspect ratio-dependent)	See 51, 83	Thermo and/or pH sensitive coating, clustering	46, 47, 84
QSY 21 (FRET acceptor)	661	> 90k	High chemical stability, resistance to photobleaching	85–87
Phthalocyanines/Porphyrins/Polypyrroles		>100k		26
Large cyclic porphyrins	746, 773, 892,939, 953	88k, 64k, 110k, 100k, 110k	Organic soluble	71
Polypyrroles	Broad 800 – 2500	See 88, 89	Hydrophobic	45, 56, 63–65
Complexed bis-metal porphyrins	360–1210 depending on chemical modifications	-	Organic soluble	72
Dendrimer phthalocyanines	685 nm	-	Water soluble	70

Fluorophore	λEX (nm)	λEM (nm)	ε (M−1 cm−1)	QY* (%)	Chemical Properties	References
Cyanines			>200k			26
Indocyanine Green (ICG)	775	831	113k#	1.3%	Water soluble but aggregates in water, binds to serum proteins. Poor photostability	34, 35 38, 40, 90–92
Cy5	648	666	250k	18%	Water soluble	19, 85, 24, 93
Cy5.5	679	696	250k	24%	Water soluble	19, 85, 93, 94
Cy7	745	775	250k	28%	Water soluble	19, 85, 24, 93, 95, 96
DiD	648	669	>125k	33%	Hydrophobic with ionic groups so interfacially active	85, 97, 98
DiR	750	782	>125k	28%	Hydrophobic with ionic side groups so interfacially active	85, 97, 98
IRDye 800CW	778	794	-	-	Water soluble	LI-COR Biosciences Inc., Lincoln NE
IR Dye 680RD/LT	680	694	-	-	RD – water soluble, optimized for small animal imaging LT – water soluble, not for small animal imaging	LI-COR Biosciences Inc., Lincoln NE
IR Dye 750	766	776	-	-	Water soluble	LI-COR Biosciences Inc., Lincoln NE
IR Dye 800RS	770	786	-	-	Water soluble nucleic acid label, not salt tolerant	LI-COR Biosciences Inc., Lincoln NE
IRDye 650	651	668	-	-	Water soluble	LI-COR Biosciences Inc., Lincoln NE
Heptamethine 3H-indolenine cyanine dyes	782–786	807–814	220k– 250k	10– 15%	Good photostability	99
Large Stokes shift cyanine derivatives	602, 617, 783	757, 803	50k, 70k, 200k	47, 38, 17%	-	100
Bis(heptamethi ne cyanine dyes)	780	800	-	-	-	101, 102
Heptamethine dyes with robust C-C bond at center	770, 800	785, 811	220k– 240k	8.8– 10%	Water soluble, monofunctional (carboxyl group)	103
Alkyl-thioether derivatized cyanine dyes	777–823	812–847	116k– 174k	2%	Water soluble, monofunctional (carboxyl group)	104
Squaraines			100k– 300k			26
Sulfonated squaraine derivatives	737–780	751–820	~200k	8–44%†	Hydro-philic and -phobic derivatives, photostable, conjugation ready	105, 106
Bis-squaraines with pyrene/thiophen e linker	Thiophene : 695, 727 Pyrene: 636, 647	Thiophene : 750, 790 Pyrene: 757, 763	Thiop hene: 74k– 240k Pyrene : 68k– 100k	0.01% - 1%	Water soluble	107
Tetralactam encapsulation of squaraines	631–661	650–704	-	8– 74%†	Polar and nonpolar derivatives studied	108
BODIPYs			>200k			26
Monomeric and polymeric BODIPY derivatives	Monomer: 597, 665 Polymer: 634–738	Monomer: 631, 701 Polymer: 669–760	-	Mono mer: 2.6, 4% Polym er: 1.1– 13%	Organic soluble dyes, good thermal and photostability	109
Heteroaryl fused BODIPYs	723, 509– 690	738, 517– 701	140k– 316k	56%, 81– 98%	Organic soluble dyes, good photostability	110, 111
BF2-chelated tetraarylazadipy rromethenes	690–706	714–730	75k– 95k	22– 30%	Both partial and fully water soluble derivatives	112, 113
BODIPY 650/665	647	664	102k	46%	Nonpolar	85, 114
Other
Alexa Fluor 647	650	668	270k	33%	Water soluble	85, 115
Alexa Fluor 660	664	691	130k	37%	Water soluble	85, 115
Alexa Fluor 680	680	704	180k	36%	Water soluble	85, 115
Alexa Fluor 700	694	720	190k	25%	Water soluble	85, 115
Alexa Fluor 750	752	776	240k	12%	Water soluble	85, 115
Alexa Fluor 790	784	814	260k	-	Water soluble	85, 116
Nile Blue	630	660	77k	27%†	Water soluble	85, 117
CellMask Deep Red Plasma Membrane Stain	649	666	-	-	Lipid soluble	85
FluoSpheres Dark Red	657	683	-	-	Carboxylate, sulfate, aldehyde, amine surfaces available	85
FluoSpheres Infrared	715	755	-	-	Carboxylate, sulfate, aldehyde, amine surfaces available	85
FocalCheck Double FarRed	669	693	-	-	Water soluble	85
HCS CellMask DeepRed Stain	648	670	-	-	Water soluble	85
HCS NuclearMask DeepRed	Broad 638	Broad 685	-	-	Water soluble	85
LIVE/DEAD Fixable NIR Cell Stain	753	776	-	-	Water soluble	85
MitoTracker DeepRed	640	662	-	-	Mitochondria stain	85
Qnuclear Deep Red Stain	642	656	-	-	Water soluble	85
SYTO 60	649	681	>50k	16%	Water soluble	85
SYTOX	641	658	-	-	Water soluble	85
TetraSpeck Dark Red Dye	656	684	-	-	Water soluble	85
TO-PRO-3	642	657	-	-	Water soluble	85
TOTO-3	643	660	-	-	Water soluble	85
Vybrant DyeCycle Ruby	638	Broad 686	-	-	Water soluble	85
DRAQ5, DRAQ7	647	Broad 665–800	21k	0.3– 0.4%	DNA dye - LIVE/FIXED (DRAQ5), DEAD/FIXED (DRAQ7). Low photobleaching.	Biostatus Ltd., Leicestershi re, United Kingdom,118 , 119
X-Sight 650 Nanospheres‡	650	673	-	-	Biocompatible, antibody conjugatable	Carestream Health, Inc., Rochester, NY
X-Sight 691 Nanospheres‡	691	715	-	-	Biocompatible, antibody conjugatable	Carestream Health, Inc., Rochester, NY
X-Sight 761 Nanospheres‡	761	789	-	-	Biocompatible, antibody conjugatable	Carestream Health, Inc., Rochester, NY
X-Sight 670 LSS Dye‡	669	755	-	-	Biocompatible, antibody conjugatable	Carestream Health, Inc., Rochester, NY
Organic soluble perylenetetracar boxdiimide derivatives	604–712	636–768	41k– 59k	NR	Organic soluble, average photostability	120
Water soluble perylenetetracar boxdiimide derivatives	432–566	588–619	10k– 33k	49– 66%	Water soluble, good photostability, nonideal emission wavelength	121
Fluorescent protein iRFP	690	713	85k– 105k	5.9%	Low photostability	122
Fluorescent protein IFP1.4	684	708	>90k	7%	Low photostability	123
Upconverting phosphors	980	540, 655	-	N/A78	Photostable, nonquenching	37, 79, 80
Quantum dots	Broad visible range	535–850	500k– 2M	60%	Photostable, nonquenching	44, 73, 77
Zn2+ multiporphyrin	Tunable: 410, 500, 690–800	Tunable: 697–817	23k– 126k	NR	Depends on side-chains, both hydrophobic	39
IRDye 700DX	680	687	170k	14%	High photostability, water soluble, acid sensitive	LI-COR Biosciences Inc., Lincoln, NE, 24
Single wall carbon nanotubes	Broad 600–800	Broad 950–1300	See 124, 125	0.01 – 0.1%	Dispersible in water with Pluronics surfactant, very photostable	125–130

^*QY = quantum yield

^#ICG’s extinction coefficient is highly dependent on concentration and solvent. Here we chose 5 mg/L in water.

^†solvent polarity dependent

^‡The X-Sight line of fluorescent spheres has been discontinued but have been included here for reference.

NR = Not reported

Cyanines

To date, the only FDA approved long wavelength dye for direct administration in medical diagnostics is indocyanine green.^{19, 132} It is one of the dyes in the cyanine family (see Fig. 10), and can be used as a fluorescent agent (see Table 1) or a NIR absorption agent.¹³³ Cyanines are characterized by two aromatic nitrogen containing heterocycles connected by a polymethine bridge. This length of the polymethine bridge is the key to tuning the excitation/emission profile of the dye. For example, Cy3 (3 methine protons) emits visible light while Cy5 (5 methine protons) and Cy7 (7 methine protons) emit in the far-red and NIR regions.²⁷ Cyanines tend to be strong absorbers with molar extinction coefficients greater than 200,000 M⁻¹ cm⁻¹ but weak fluorescers with quantum yields from 1 – 18%. However, recent work has yielded positive exceptions to this generalization, as seen in Peng, et al.¹⁰⁰ Extending the polymethine bridge tends to lower quantum yields but shifts emission into the NIR.^{19, 26} In addition, it has been found that the addition of the cyclohexenyl moiety at the center of the polymethine bridge as well as complexation with proteins tends to raise the quantum yield and stability of the dyes.^{26, 101} This effect is generally attributed to ‘rigidizing’ the backbone.¹⁰² On the other hand, Chen, et al. substituted various electron donating or withdrawing groups at the N position on the 3H-indolenine rings (rather than modifying the polymethine bridge). They found that electron donating groups offered greater photostability to the molecule while withdrawing groups decreased stability.⁹⁹

Fig. 10.

Clockwise, from left to right, (adapted) examples of cyanine dyes ICG⁹⁰, Cy5¹³¹, and IRDye 750 NHS Ester (adapted from LI-COR Biosciences Inc., Lincoln NE).

Peng, et al.¹⁰⁰ synthesized novel heptamethine cyanine dyes through modification of the central chlorocyclohexenyl group resulting in larger Stokes shifts (>140 nm) and higher fluorescence. However, the dyes were significantly blue shifted with lower molar extinctions but higher quantum yields. Lee, et al. have also developed a novel series of cyanine dyes by reacting the same group to add a carboxyl containing moiety to the central cyclohexenyl group, thus improving water solubility and offering a site for bioconjugation.¹⁰³ Their dyes, in contrast, were red-shifted (both absorption and emission) with lower quantum yields.

Squaraines

Squaraines are a class of dyes consisting of an oxocyclobutenolate core with aromatic or heterocyclic components at both ends of each molecule (see Fig. 11).²⁶ They are characterized by high extinction coefficients, quantum yields, and photostabilities.¹⁰⁵ However, their planar, highly conjugated architecture results in dyes with poor water solubility. Thus, improving water solubility is an important goal for many researchers in this area. Umezawa, et al. synthesized sulfonated squaraine derivatives with excellent water solubility and emission wavelengths above 800 nm. In addition, they found that adsorption onto BSA proteins greatly increased the fluorescent output, similar to the cyanines. However, the process of sulfonation and the presence of highly polar water reduced the quantum yield to 8% whereas similar nonpolar compounds had quantum yields >40% in organic solvents.¹⁰⁵ Water soluble squaraines were also synthesized by Nakazumi, et al. by forming bis-squaraines separated by pyrene or thiophene linkers and carboxyl end groups. They, too, found that the fluorescence and stability of squaraines improved when bound to HSA or BSA. However, quantum yields were very low, on the order 0.01 to 1% and emission wavelengths were 750–790 nm.¹⁰⁷

Fig. 11.

Examples of squaraines (adapted from Umezawa¹⁰⁵ and Nakazumi¹⁰⁷).

Gassensmith, et al. have taken a different approach to improving water solubility through the encapsulation of squaraines within tetralactam macrocycles. This has the dual effect of preventing aggregation in aqueous solution and protecting the squaraines from chemical attack though they are relatively stable species even without encapsulation. Encapsulated species showed emission bands from 650–700 nm with relatively high quantum yields in nonpolar solvents which decreased in the presence of water or other polar solvents. In vivo imaging was demonstrated by injected squaraine-labeled E. coli and S. aureus into live nude mice and observing the fluorescent emission.¹⁰⁸

BODIPYs

BODIPY (borondipyrromethene) class dyes were originally discovered in 1968 by Treibs and Kreuzer¹³⁴ and have received much attention as agents for biomedical imaging (see Fig. 12). They generally have sharp absorbance and emission profiles and high quantum yields approaching 100%, and have solvent and pH insensitive emission profiles with good stability under physiological conditions.²⁵ However, they have low extinction coefficients (80,000 M⁻¹ cm⁻¹) and very few are water soluble.²⁶ Furthermore, most of them emit below 600 nm, making them unsuitable for deep tissue imaging.²⁵ Small chemical modifications to the basic BODIPY structure can cause large shifts in the emission profile as well as improving water solubility and the extinction coefficient. Donuru, et al. have taken the approach of synthesizing novel polymeric and copolymeric BODIPY dyes, tuning the emission properties by introducing styryl groups, thus extending the π conjugated system. These hydrophobic polymeric dyes emit from 669 – 760 nm with quantum yields from 1.1 – 13%.¹⁰⁹

Fig. 12.

Examples of BODIPY class dyes (adapted from Donuru¹⁰⁹ and Umezawa¹¹⁰).

The Suzuki group has taken a variety of approaches, synthesizing a whole range of BODIPY based monomers with a variety of side groups attached to the core BODIPY moiety. The first set of dyes were heteroaryl-fused BODIPYs with tuned emission wavelengths from 583–738 nm and greatly improved extinction coefficients above 185,000 M⁻¹ cm⁻¹.¹¹¹ In 2009, this strategy was extended to the addition of several types of electron donating moieties as well as the synthesis of asymmetric BODIPY derivatives. The new set of dyes had emission wavelengths as high as 701 nm with high quantum yields and extinction coefficients as high as 300,000 M⁻¹ cm⁻¹.¹¹⁰ These dyes exhibited excellent photostability but do not appear to be water soluble.

Tasior and O’Shea, on the other hand, have synthesized water soluble and partially water soluble BODIPY based dyes with easily conjugated side groups.^{112, 113} The main strategies involved attaching sulfonate or carboxylate terminated alkyl chains or the introduction of quaternary ammonium salts to the molecule to impart charge. These compounds were red shifted into the NIR range with an emission range of 718–730 nm in phosphate buffered saline with bovine serum albumin. Extinction coefficients and quantum yields were lower across the board than the non-water soluble BODIPYs, ranging from 51k–89k M⁻¹ cm⁻¹ and 22 – 31%, respectively. Nonetheless, these studies represent a step in a positive direction for biomedical imaging.

Perylene Derivatives

Falling somewhat outside the major categories stated above, Quante, et al. synthesized a set of perylenetetracarboxdiimide derivatives with highly extended π-conjugated systems (see Fig. 13). This class of dyes is generally characterized by excellent chemical, thermal and photostability^{121, 135} but has limited solubility in both organic and aqueous solvents and somewhat low extinction coefficients (50,000 M⁻¹ cm⁻¹). Red-shifting was primarily accomplished by extension of the π system as well as the introduction of halogens to induce bathochromic shifts. The resulting dyes exhibited long wavelength emission in the range 636 – 768 nm and improved organic solubility without a major loss of stability. Quantum yields were not, however, reported.¹²⁰ For aqueous applications, Qu, et al. reported the synthesis of water soluble perylenetetracarboxdiimide derivatives by introducing charged groups such as sulfonates. The ionic dyes showed good aqueous solubility and photostability with quantum yields in excess of 50%. However, extinction coefficients were low (10,000 – 30,000 M⁻¹ cm⁻¹) and emission wavelengths reached maximum of only 619 nm, outside of the NIR range.¹²¹

Fig. 13.

Perylene and an example derivative (adapted from Quante, et al.)¹²⁰

Carbon Nanotubes

A rather unexpected development involves the possibility of using single walled carbon nanotubes (SWCNT) as NIR fluorescence agents. The Weisman lab conducted numerous fundamental as well as biological studies involving these materials.^{126, 128, 129} The nanotubes the studies tend to be on the order of nanometers in diameter but 100 nm in length and exhibit number of excitation bands from 600 – 800 nm and emission bands from 950 – 1300 nm. Their quantum yields were measured by Huang, et al. and found to be 0.01 – 0.1%, orders of magnitude lower than many organic fluorophores. Despite this limitation, however, they report that similar signal to noise ratios were achieved due to a high absorption coefficient.¹²⁷ In Cherukuri, et al., pristine hydrophobic SWCNTs 1 nm in diameter and 1µm in length were taken up into phagocytic cells at approximately 1 nanotube per second per cell (Fig. 14). They were found to be photostable after 100 min of tracking and did not cause any cytotoxic effects. However, the nanotubes were stabilized by Pluronics surfactant in order to prevent aggregation in vitro. It is not clear whether that strategy would be efficient in vivo.¹²⁶ The use of functionalized nanotubes has been reported by Pantarotto, et al. for the delivery of DNA to HeLa cells. Their 20 nm diameter, 200 nm length nanotubes were functionalized with ammonium terminated oligoethylene glycol chains imparting an overall positive charge and making them water soluble. These nanotubes were complexed with negatively charged plasmid DNA which led to the formation of supramolecular structures such as spheres, toroids or supercoils. The nanotubes were readily taken up into HeLa cells and 5 – 10 fold increases in gene expression were observed compared to DNA treatment alone. Undoubtedly, the cationic charge facilitated nanotube penetration as the delivery mechanism was determined to be passive, not active.¹³⁰

Fig. 14.

Carbon nanotube (CNT) fluorescence emission after uptake into a macrophage-like cell (reproduced from Cherukuri, et al.)¹²⁶

Fluorescent Proteins

The development of Green Fluorescent Protein (GFP) revolutionized the study of diseases, cancer in particular, providing insights into real-time growth and metastatic behaviors.²³ In recent years, development of fluorescent proteins has proceeded to the point that researchers now have a variety of proteins as molecular probes spanning the visible spectrum.^136–139 For whole animal imaging applications, bacterial phytochromes provide the best scaffold for NIR excitation and emission wavelengths, as opposed to a GFP-like scaffold.¹⁴⁰ Phytochromes are photosensory receptors that absorb light at far-red and near-infrared wavelengths. According to Shu, et al., Rhodopseudomonas palustris and Pseudomonas aeruginosa bacteriophytochromes expressed in E. coli have been found to emit photons from 710 to 725 nm. Their engineered proteins derived from Deinococcus radiodurans bacteriophytochromes excite at 684 nm and emit at 708 nm with >90k M⁻¹ cm⁻¹ extinction coefficient and 7% quantum yield.¹²³ Filonov, et al. have engineered NIR fluorescent proteins from the Rhodopseudomonas palustris phytochrome basis. The best variant they produced has maxima at 690 nm (ex) and 713 (em) and an aqueous quantum yield of 5.9%. The photostability of the protein is rather low, exhibiting a 450 s half life under photobleaching light irradiation. However, it appears to have a higher brightness than the constructs of Shu, et al.¹²² Both proteins are reliant on cofactors (such as biliverdin IXα, readily available as an intermediate in mammalian heme metabolism) for maximum emission.¹²² Clearly, there is much promise and as yet untapped potential for the development of long wavelength protein fluorophores. The ability of these agents to be inserted into genomic or plasmid DNA sequences to monitor cell metabolism and expression pathways makes them invaluable for biomedical research.

IV. Future Perspectives: Multifunctional NPs for Diagnostics and Delivery

While the previous sections focused on the development of fluorophores, we now turn our attention to the means by which fluorophores can be delivered or targeted for maximum efficacy. When the goal is diagnostic or fundamental biology, it is often enough to have a targeting agent (ligand, antibody, etc.) directly attached to the fluorophore or contrast agent. That has driven the effort to synthesize aqueous-soluble fluorophores, as has been mentioned above. However, there are significant advantages for diagnostics and imaging to place the imaging agent in a nanocarrier format. The larger sizes of nanocarriers increase circulation times, if they are sterically well protected. Increased circulation time translates to better targeting in vivo and in vitro. Most importantly, a single 100 nm particle may carry 15,000 dye molecules, therefore brightness per particle is greatly enhanced over single molecule delivery. The location of the imaging agent in the nanocarrier influences brightness. If the imaging agent is on the surface of the particle, increasing size decreases the brightness per mass of imaging agent since the surface-to-volume ratio decreases with increasing size. Therefore, nanoparticles with dye localized in a hydrophobic core have several advantages: (1) the possibility of increased brightness due to the core volume increasing linearly with mass, (2) a large variety of dyes that cannot be made in hydrophilic form, or which are brighter in hydrophobic form, (3) decreased interactions with the biological environment and less sensitivity of emission to environment, and (4) increased stability against photobleaching or oxidative degradation.

For therapeutic applications, the requirement of multifunctionality can involve delivery of an active agent, targeting, and visualization of delivery. This generally requires the construction of nano-carriers that incorporate these functions. Completely soluble nanocarrier systems include linear polymer chains with grafted pro-drugs, targeting ligands, and imaging agents.¹⁴¹ Dendrimers, highly branched polymeric structures, have been attractive multifunctional nanocarriers because their high degree of functionality makes multifunctional conjugation possible. However, sequential functionalization of soluble nanocarriers has made precise quantification of the functionalization of each carrier (as opposed to the average composition) difficult. This has been a stumbling block to regulatory approval for some of these materials.

Insoluble nanocarriers have some advantages in more precise control of composition. They also offer particular advantages for NIR imaging in that the hydrophobic cores of nanocarriers enable the encapsulation of hydrophobic imaging agents. The greater rigidity of the environment increases quantum efficiency, decreases photobleaching,³⁶ and decreases quenching.^{36, 142} The insoluble carriers involve either self-assembled micelle structures, or precipitated nanostructures, or carriers made by emulsification and then solvent stripping.

Silica Nanocarriers

Benezra, et al. have developed novel nanoparticles that encapsulate fluorescent dyes within an amorphous silica shell that is coated with PEG for increased circulation time as well as ¹²⁴ I-cRGDY-PEG to allow for multimodal imaging and α_vβ₃ integrin targeting. The core of the nanoparticle contains the dye Cy5 (see Table 1) while the ¹²⁴ I allows for a second imaging modality through PET. These particles are about 7 nm in diameter and thus very small compared to most liposomes and polymeric nanoparticles. This small size allows for fast renal clearance and wide biodistribution but the presence of PEG and targeting moieties increase circulation time to give a blood half-life of 5.6 hours while excretion analysis showed that ~50% of the particles are eliminated in the first 24 hours in mouse tumor models. Though a longer wavelength fluorescent dye would have improved the penetration depth, Cy5 fluorescence imaging allowed the authors to elucidate sub-millimeter nodal features that PET could not, though PET may be better for deeper penetration and quantitation. These particles have been given FDA IND approval for clinical trials.¹⁴³

Micellar Nanocarriers

Micellar nanocarriers involve an amphiphilic block copolymer which defines the nanocarrier size, the hydrophobic micelle core where drugs and fluorophores can be captured, and the hydrophilic corona which can confer biocompatibility and targeting.^{144, 145}

A particularly interesting imaging agent was developed by Rodriguez, et al. consisting of ICG complexed to quaternary ammonium salts to render them hydrophobic, and enabling encapsulation within 30 nm block copolymer micelles (Fig. 15). This process improved the thermal- and photo-stability of the ICG with a modest red shift in the spectra. Encapsulation within micelles also has the potential to greatly improve in vivo circulation times by preventing plasma protein binding, a key weakness of free ICG. Lastly, this method allows for targeting of the micelles through protein or small molecule conjugation to the hydrophilic block of the copolymer micelles.⁴⁰

Fig. 15.

Atomic force microscope (AFM) images of 30 nm micelles encapsulating hydrophobic ICG complex. Scale bar applies to both images. Reproduced from Rodriguez, et al.⁴⁰

Ghoroghchian, et al. have constructed 50 nm – 20 µm size polymersomes built from self-assembled poly(butadiene)-b-poly(ethylene oxide) (PBD-b-PEO) amphiphilic block copolymers. The Zn²⁺-porphyrin fluorophores varied in macrocycle-macrocycle linkage topology, allowing for fine tuning of the emission from 600 – 825 nm. These compounds were stably incorporated into the polymersome membranes, protecting them from chemical attack and potentially improving circulation times in vivo. They were found to be well distributed throughout the membrane without phase separation though modulation of the hydrophobicity of the porphyrin complexes moved the complexes from the hydrophobic interior to the PBD-PEO interface.³⁹ The quantum yield is expected to be high given the yields of similar compounds discussed earlier.

Precipitated Nanoparticle Carriers

Akbulut, et al. utilized the versatile Flash NanoPrecipitation (FNP) process to encapsulate a variety of fluorophores (pyrene, anthracene, nile red, porphirine) with emissions ranging from 370 to 720 nm. Briefly, this mechanism of nanoparticle formation (Fig. 16) involves the intense micromixing of an organic stream containing a dissolved hydrophobic fluorophore and amphiphilic block copolymer such as poly(caprolactone)-b-poly(ethylene glycol) (PCL-b-PEG) against a large excess of non-solvent such as water. The resulting drop in solvent quality causes rapid precipitation of the fluorophore while the rapid mixing time ensures uniform growth kinetics. Tuning of the nucleation and growth time scale against the copolymer aggregation time scale ensures that the block copolymer anchors (i.e. adsorbs) to the hydrophobic core surface at the right time to allow nanoparticle production from 30 – 800 nm in size with narrow size distributions.^{36, 146} In addition to fluorescent molecules, this process has been applied successfully to encapsulate a variety of drugs, inorganics, and peptides.^{37, 142, 148–153} The stealth PEG coating improves circulation time in vivo and provides steric protection against aggregation.¹⁴⁸ The fluorescent nanoparticles in this study were found to be bright with increased photostability over free dye, probably due to molecule immobilization and copolymer protection. The copolymer chains can be conjugated to targeting ligands such as antibodies¹⁵⁰ to provide targeting capabilities. Additionally, Gindy, et al. showed that more than one type of solute can be encapsulated within these nanoparticles cores, thus the possibility of diagnostic and therapeutic functions in one package – theranostics.¹⁵¹ The main requirement for the fluorophore in this process is that they must be hydrophobic for maximum stability. If they are not, chemical modification is often necessary, as demonstrated by Akbulut, et al.^{36, 148}

Fig. 16.

Flash NanoPrecipitation schematic. Rapid mixing of the solvent and nonsolvent streams causes a drop in solvent quality and subsequent precipitation of the hydrophobic solutes. By matching the aggregation (t_agg) and nucleation and growth (t_ng) time scales (which are both larger than the mixing time scale t_mix), homogeneous nucleation kinetics result and polymer stabilized nanoparticles from 30 – 800 nm with narrow size distributions are produced. This scalable method allows for rapid, inexpensive, and versatile encapsulation of various hydrophobic molecules.^{146, 147}

As mentioned earlier, Budijono, et al. and Shan, et al. have demonstrated the feasibility of encapsulating inorganic upconverting phosphors alongside tetraphenylporphyrin photosensitizer using the Flash NanoPrecipitation process (Fig. 17). The ~200 nm particles produced were stabilized with poly(lactic acid)-b-poly(ethylene oxide) and found to be stable under physiological conditions. More importantly, they demonstrated strong cancer cell killing activity from the production of reactive oxygen species (ROS) under 980 nm illumination. While the UCP emission band is largely absorbed by the porphyrin photosensitizer, fluorescence from the UCPs can be detected for diagnostic purposes making this system multifunctional, though targeting studies have not yet been conducted.^{37, 79}

Fig. 17.

SEM image of poly(lactic-co-glycolic acid)-b-poly(ethylene glycol) protected upconverting nanophosphors (140 nm) with additional 30 nm micelles visible. Reproduced from Budijono, et al.³⁷

In addition to UCP-photosensitizer pairing to enable deep tissue ROS generation, another hybrid system involves the pairing of gold structures with NIR emitting fluorophores such as indocyanine green (ICG). As reported by Tam, et al. and previously discussed, gold nanoparticles exhibit a size- and morphology-dependent surface plasmon resonance absorption band centered at 520 nm for 5 nm spherical gold nanoparticles. Increasing the size of these spheres or using nanorods or shells can shift the absorption band into the NIR region. Placing a molecular fluorophore such as ICG directly onto a metal surface generally results in quenching. However, maintaining a few nanometers separation and matching the plasmon resonance of ~120 nm gold nanoshells with the emission band of the ICG, the fluorescence was enhanced by a factor as high as 50, greatly improving the effective quantum yield.^{52, 154} This effect stems primarily from the effect of an additional radiative decay rate introduced by the metal surface:

$$Q_0=\frac{\mathrm{\Gamma }}{\mathrm{\Gamma }+k_{\mathit{\text{nr}}}}$$ (1)

$$Q_m=\frac{\mathrm{\Gamma }+{\mathrm{\Gamma }}_m}{\mathrm{\Gamma }+{\mathrm{\Gamma }}_m+k_{\mathit{\text{nr}}}}$$ (2)

• Q₀, Q_m – original, modified quantum yield

• Γ – radiative decay rate

• Γ_m – metal induced radiative decay rate

• k_nr – nonradiative decay rate

This additional radiative decay Γ_m in Eq. 2 (compared to Eq. 1) increases the rate of resonance energy transfer and improves the quantum yield.¹⁵⁵ This effect was exploited by Chen, et al. to formulate a multifunctional theranostic system for ovarian cancer diagnosis and treatment (Fig. 18). Briefly, a ~70 nm gold nanoshell is encapsulated in a silica shell which is doped with superparamagnetic iron oxide (SPIO) and ICG. This system benefits from the gold-enhanced fluorescence of ICG, an additional MRI imaging mode using SPIOs, and a gold shell to convert absorbed light to heat for photothermal cancer cell ablation. Finally, anti-HER2 antibodies were chemically conjugated to the nanocomplexes for targeting functionality. These particles were found to be bright and effective in reducing populations of OVCAR3 ovarian cancer cells in vitro through the photothermal effect after being taken up via HER2 receptor-mediated routes.³⁸ Though in vivo results were not shown, these theranostic complexes seem very promising as the next generation in pharmaceutical paradigms.

Fig. 18.

Novel theranostic agent. A – schematic of nanoparticles containing a ~70 nm Au nanoshell with a silica shell doped with superparamagnetic iron oxide and ICG and surface-decorated with anti-HER2 antibodies for targeting. B – magnetic resonance imaging of the nanoparticles in vitro (no scale bar provided). C – photothermal ablation capabilities of the theranostic system demonstrated in vitro. D – fluorescence visualization of the nanoparticles after targeted uptake into OVCAR3 cells. Reproduced from Chen, et al.³⁸

V. Conclusions

Combining increasingly bright and stable NIR fluorophores with effective delivery systems that enhance the fluorescence, circulation time, and photochemical stability of encapsulated fluorescent molecules is already providing powerful new tools for fundamental biological studies as well as diagnostic applications. With the addition of drug payloads and targeting capabilities, multifunctional ‘theranostic’ agents will be able to provide a complete solution for effective diagnostics and disease therapy. These agents will combine the best of what synthetic organic chemistry, pharmaceutical- and bio-engineering have to offer to create a new paradigm in the biomedical sciences.