2.1.

Cell Lines

The targeting of EGFR- and CyclinD1-specific optical contrast agents was evaluated using 1483 cells and xenograft tumors. This EGFR-positive cell line, derived from a patient with head and neck squamous cell carcinoma,⁴² was obtained from Dr. Reuben Lotan at the M.D. Anderson Cancer Center (Houston, Texas). The 1483 cells were cultured in Dulbecco’s Modified Eagle Medium: Nutrient Mix F-12® medium supplemented with L-glutamine (Invitrogen, Carlsbad, California) and 10% fetal bovine serum (Hyclone, Logan, Utah).

2.2.

Tissue Models

Three different tissue models were evaluated for permeability in the presence and absence of topical permeation enhancers. Porcine oral mucosa was selected as a model of stratified squamous epithelium. Heads of American Yorkshire pigs, aged $6\text{months}$ , were obtained from a local slaughterhouse at the time of sacrifice. The buccal mucosa of the oral cavity and $\sim 1\mathrm{mm}$ of underlying tissue was separated from the sur-rounding muscle layers via dissection. Guinea pig bladder mucosa was selected as a model of transitional epithelium. Whole bladders were excised from $2–3\text{week}$ old female Hartley guinea pigs (Charles River Laboratories, Wilmington, Massachusetts) directly following animal sacrifice. Xenograft tumors, generated subcutaneously in mice, were selected as a model of squamous cell carcinoma for the study of cancer biomarker targeting. Briefly, 1483 cells ( $2\times {10}^6$ viable cells in $100\mu \mathrm{L}$ PBS) were implanted subcutaneously in the left and right posterior mammary fat pads of $6–8\text{week}$ -old female Nu/Nu mice (Charles River Laboratories). When tumors reached $4–5\mathrm{mm}$ diam, the animals were anaesthetized and sacrificed via cervical dislocation. All animals were cared for in accordance with institutional guidelines. The protocols were reviewed and approved by the Institutional Animal Care and Use Committee at Rice University.

2.3.

Permeation of Fresh Tissues

To reproducibly permeate tissues for contrast-agent delivery, tissue biopsies of uniform surface area were topically treated with permeation enhancers. A $4\text{-}\mathrm{mm}$ -diam dermal punch (Miltex Inc., York, Pennsylvania) was used to produce cylindrical samples of oral and bladder mucosa. Subcutaneously generated tumors, grown as cylinders to the appropriate diameter, were cut in half to expose the tumor surface. The tissue cylinders were embedded vertically into 3% (vol/vol) ultrapure agarose (Invitrogen) to prevent the influx of permeation enhancers and contrast agents at the tissue margins. The Darcy permeability of 3% agarose has been reported as $\sim 450{\mathrm{nm}}^2$ ,⁴³ and the diffusivity of macromolecules through 3% agarose is limited.⁴⁴ The apical surface of the epithelium was left free of agarose to facilitate topical labeling. In the case of the bladder tissue, the underlying muscle layers were sufficiently thick to allow the biopsies to be positioned horizontally in the agarose without stretching or bending. Two topical permeation enhancers, Triton-X100 and DMSO (Sigma-Aldrich, St. Louis, Missouri) diluted in phosphate buffered saline [(PBS), Sigma-Aldrich], were evaluated for their ability to permeate fresh tissue. PBS alone was used as a negative control. One milliliter of permeation-enhancing solution was topically applied to the apical surface each tissue biopsy for $0–4\mathrm{h}$ at $4°\mathrm{C}$ .

2.4.

Permeabilization Detection Using Macromolecules

To determine whether Triton-X100–treated tissues are selectively permeable to macromolecules of a specific size, bladder biopsies were probed with fluorescent macromolecules of three sizes. The apical surface of the epithelium was topically treated for $1\mathrm{h}$ with 0.5% Triton-X100 (to yield a treatment dose of $\sim 0.8\mathrm{pmol}$ /cell, assuming full-thickness permeation) or PBS, washed once in PBS, and covered with a 1:1:1 mixture of rhodamine-dextran ( $3\mathrm{kDa}$ , Invitrogen), fluorescein-dextran ( $40\mathrm{kDa}$ , Invitrogen), and AlexaFluor647 IgG ( $150\mathrm{kDa}$ , Invitrogen), each diluted to a concentration of $1\mu \mathrm{M}$ in PBS. Tissues were immersed with this solution for $15\min$ at $4°\mathrm{C}$ and then imaged using fluorescence confocal microscopy at three different excitation wavelengths (described below). Images were collected in $5\text{-}\mu \mathrm{m}$ steps from the surface into the tissue. Following imaging, the tissues werewashed three times in cold PBS ( $15\min$ total) and reimaged to assess the removal of unbound macromolecules. A total of 10 bladders were evaluated in independent experiments.

2.5.

Tracking the Time Course of Tissue Permeation

To monitor Triton-mediated permeation of bladder epithelium as a function of time, fresh biopsies were treated for different time intervals. The apical surface of the agarose-embedded punch biopsies were topically treated with 0.5% Triton-X100 for 0, 5, 10, 15, 30, or $60\min$ . Following the permeation treatment, the tissues were washed once in cold PBS and probed with the 1:1:1 mixture of fluorescent macromolecules described above. Confocal microscopy images were collected $40\mu \mathrm{m}$ below the tissue surface to allow for optical sectioning of both the epithelium and underlying tissues. Each time point was evaluated in five independent experiments.

2.6.

Determination of the Depth of Penetration

The depth of Triton-mediated macromolecule penetration as a function of time and concentration was evaluated in cross sections of fresh oral mucosa. The apical surface of agarose-embedded punch biopsies were topically treated for $0–4\mathrm{h}$ with Triton-X100 concentrations ranging from 0 to 2.5%. These concentrations were selected to treat cells with up to $1.1\mathrm{pmol}$ /cell of Triton-X100. The permeabilized tissues were cross-sectioned using a Krumdieck tissue slicer (Alabama Research & Development, Munford, Alabama) and immersed for $15\min$ in the 1:1:1 mixture of fluorescent macromolecules. Confocal fluorescence images of the transverse sections were collected $40\mu \mathrm{m}$ below the cut surface to avoid confounding effects due to damage at cut surface. The penetration depth for each size of macromolecule was evaluated by measuring the distance between the apical surface of the tissue and the leading edge of the fluorescent macromolecules using ImageJ v1.34 software (http://rsbweb.nih.gov/ij). Measurements were collected at $20\text{-}\mu \mathrm{m}$ intervals across four representative images in four independent experiments (16 images total per treatment condition). The rate of tissue permeation as a function of Triton-X100 concentration was determined from a linear least-squares fit to the data.

2.7.

Comparison of Macromolecule Penetration in Normal and Cancer Tissue

To determine whether the depth and rate of permeation varies between tissues, macromolecule penetration was assessed in fresh normal oral mucosa and squamous cell carcinoma. The apical surface of agarose-embedded oral biopsies and 1483 tumors were treated for $1\mathrm{h}$ with only 1% Triton-X100, 10% DMSO, or PBS. Following the permeation treatment, the tissues were cross-sectioned with a Krumdieck tissue slicer and probed with the 1:1:1 mixture of fluorescent macromolecules for $15\min$ . Confocal fluorescence images of the transverse sections were collected $40\mu \mathrm{m}$ below the cut surface. The depth of penetration from the apical surface was assessed for each macromolecule at $20\text{-}\mu \mathrm{m}$ intervals across four representative images from four independent experiments (16 images total per treatment condition). The rate of tissue permeation was determined from a linear least-squares fit to the data.

2.8.

Estimation of the Normalized Triton-X100 Concentration in Tissue

The average dose of Triton-X100 per cell was determined as a function of treatment time and depth. Briefly, the average permeation depth was measured at different time intervals for tissues of known surface area treated with a fixed volume and concentration of Triton-X100. An average cell was approximated as a sphere $\sim 8.33\mu \mathrm{m}$ in diameter based on confocal microscopy images. The measured volume of treated tissue was divided by the average cell volume to yield the number of treated cells. The concentration of Triton-X100 (in moles per milliliter) was divided by the number of treated cells to determine the normalized Triton-X100 concentration.

2.9.

Cell Viability Assessment in Tissue Phantoms

Cell viability following Triton-X100 treatment was assessed in three-dimensional tissue phantoms. The optical properties of these phantoms have been validated in Ref. 45. Briefly, subconfluent monolayers of 1483 cells were treated with 0.05% trypsin-EDTA (Invitrogen), pelleted, and washed once with cold PBS. The cells were counted under light microscopy using a hemocytometer, pelleted, in resuspended in a collagen type I gel ( $2.1\mathrm{mg}∕\mathrm{mL}$ , Roche Diagnostics Corp, Indianapolis, Indiana) at concentration of $1–10\times {10}^7\text{cells}∕\mathrm{ml}$ . Then, $100\mu \mathrm{L}$ aliquots were cured for $30\min$ at $37°\mathrm{C}$ and then placed in prewarmed complete media, and $24\mathrm{h}$ later, the tissue phantoms were treated with $1\mathrm{mL}$ of 0–5% Triton-X100 for $1\mathrm{h}$ at $4°\mathrm{C}$ . The Triton-X100 concentrations were selected to yield a normalized concentration of $0.55\mathrm{pmol}$ /cell and above to permeabilize all cells. After treatment, the phantoms were washed and placed in fresh prewarmed media. Cell viability was assessed $4\mathrm{h}$ following Triton-X100 treatment. Prewarmed 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT, Sigma-Aldrich) was added to the supernatant for $20\min$ at a concentration of $0.5\mathrm{mg}∕\mathrm{mL}$ . In viable cells, reduction of the MTT reagent led to the intracellular deposition of a blue formazan product visible to the eye. The tissue phantoms were imaged whole and in cross section using light microscopy.

2.J.

Synthesis and Validation of Molecular-Specific Contrast Agents

Mouse antihuman antibodies targeted to epidermal growth factor receptor [(EGFR), clone 108; custom synthesized by the Baylor College of Medicine, Houston, Texas] and Cyclin D1 (clone 72-13G; Santa Cruz Biotechnology Inc., Santa Cruz, California) were reacted with AlexaFluor® 647 and AlexaFluor® 488 carboxylic acid succinimidyl esters using commercially available labeling kits (Invitrogen). The purified conjugates were suspended in PBS at a concentration of 1.0 and $0.2\mathrm{mg}∕\mathrm{mL}$ , respectively. Dye-labeled isotype controls were synthesized at the same concentrations. Prior to tissue labeling, the bioactivity and specificity of the conjugates wereconfirmed in live and fixed 1483 cells as described in Ref. 46, both in the presence and absence of Triton-X100.

2.K.

Comparison of AntiEGFR Delivery Strategies in Tumors

To assess the ability of permeation enhancers to deliver molecular-specific optical contrast agents through tissue, subcutaneously generated 1483 xenograft tumors were for labeled for EGFR in the presence of 1% Triton-X100, 10% DMSO, or PBS. AntiEGFR-647 (1:50) was diluted in permeation enhancers and topically applied to the cut surface of agarose-embedded tumors for $1\mathrm{h}$ at $4°\mathrm{C}$ . The labeled tissues were washed three times in cold PBS, sliced perpendicular to the surface using a Krumdieck tissue slicer, and counterstained for $30\mathrm{s}$ with 0.05% acriflavine-HCL (Sigma-Aldrich), a cell permeant nucleic acid dye. The depth and localization of EGFR labeling was assessed from confocal fluorescence images acquired $40\mu \mathrm{m}$ below the cross section’s surface. A total of 10 tumors (5 mice) were evaluated.

2.L.

Demonstration of Multitarget Labeling in Tumors

To demonstrate the feasibility of simultaneously targeting multiple cancer biomarkers in different cell compartments, 1483 xenograft tumors were simultaneously labeled with three tissue and/or cell impermeant probes. AntiEGFR-647 (1:50), antiCyclinD1-488 (1:10), and propidium iodide ( $15\mu \mathrm{M}$ , Sigma Aldrich) were used to label cell membrane, cytoplasmic, and nuclear targets respectively. The contrast agents and isotype controls were diluted in 1% Triton-X100 and topically applied for $1\mathrm{h}$ . The labeled tissues were washed three times in cold PBS and sliced perpendicular to the surface using a Krumdieck tissue slicer. The localization of labeling was assessed from confocal fluorescence images acquired $40\mu \mathrm{m}$ below the cross section’s surface. A total of 10 tumors (5 mice) were evaluated.

2.M.

Confocal Image Acquisition and Analysis

All images were obtained using a Carl Zeiss LSM 510 confocal microscope (Thornwood, New York) equipped with $488\mathrm{nm}$ $\left(30\mathrm{mW}\right)$ , $543\mathrm{nm}$ $\left(1\mathrm{mW}\right)$ , and $633\mathrm{nm}$ $\left(5\mathrm{mW}\right)$ lasers. Images were collected using photomultiplier tube detectors and Zeiss LSM 5 image examiner software. Samples were sequentially excited with each laser line, with power settings of 50, 100, and 100%, respectively. Fluorescence emission was collected using 500–530, 565–615, and $650–710\mathrm{nm}$ bandpass filters, respectively. Tissue reflectance at $633\mathrm{nm}$ was collected using a $635\text{-}\mathrm{nm}$ dichroic beam splitter. Images were acquired at $0.5\mathrm{fps}$ using a 63X oil, 20X, or 10X objectives with a pinhole of 2.56 Airy units. For the nonspecific macromolecule permeation studies, the gain was held constant just below saturation level of the extracellular solution. In the molecular-specific targeting studies, the gain was held constant at 440 for antiEGFR-647 ( $633\text{-}\mathrm{nm}$ excitation), 650 for antiCyclin D1-488 ( $488\text{-}\mathrm{nm}$ excitation), 400 for propidium iodide ( $543\text{-}\mathrm{nm}$ excitation), and 535 for acriflavine-HCL ( $488\text{-}\mathrm{nm}$ excitation).

3.1.

Permeabilization of Fresh Guinea Pig Bladder

The ability of Triton-X100 to permeabilize guinea pig bladder mucosa was assessed using untargeted fluorescent macromolecules of three different sizes. Figure 1 shows representative confocal images of tissue biopsies topically treated for $1\mathrm{h}$ with 0.5% Triton-X100 or PBS, and then probed with a 1:1:1 mixture $3\mathrm{kDa}$ rhodamine-dextran, $40\mathrm{kDa}$ fluorescein-dextran, and $150\mathrm{kDa}$ Alexa647-IgG. The tissue reflectance images are shown on the left and the corresponding fluorescence images are shown to the right. The yellow/white lines indicate the basal boundaries of the epithelium. Because of the three-dimensional folding of the deflated bladder, it was possible to image the transitional epithelium in cross section using confocal microscopy at a depth of $40\mu \mathrm{m}$ . In the reflectance images, the epithelium was distinguished from the underlying, more highly scattering tissues by its darker appearance. Triton-X100 treatment facilitated trans-epithelial penetration of all three sizes of molecules. The 3 and $40\mathrm{kDa}$ molecules successfully penetrated both the cytoplasm and nucleus of epithelial cells, whereas the $150\mathrm{kDa}$ molecules only penetrated the cytoplasm of epithelial cells. With several brief washes, a significant portion of the macromolecules were removed. In the PBS-treated controls, the macromolecule penetration was limited to one to two layers of superficial cells.

Fig. 1

Macromolecule penetration of guinea pig bladder epithelium following permeation treatment. Tissues were topically treated for $1\mathrm{h}$ with 0.5% Triton-X100 or PBS and then probed with a 1:1:1 mixture of $3\mathrm{kDa}$ rhodamine-dextran, $40\mathrm{kDa}$ fluorescein-dextran, and Alexa647-IgG. The yellow/white lines indicate the basal surface of the epithelium. In reflectance images, the epithelium was distinguished from the underlying tissues by its darker appearance. Following treatment with Triton-X100, trans-epithelial penetration was observed for all three sizes of molecules. After washing with PBS, the fluorescence intensity of the labeled tissue was significantly reduced. In PBS-treated controls, macromolecule penetration was limited to a few cells in the superficial layers. The scale bar represents $100\mu \mathrm{m}$ .

The time course of tissue permeation with Triton-X100 treatment was tracked using confocal microscopy. Figure 2 shows representative images of biopsies treated with 0.5% Triton-X100 for 5, 10, and $15\min$ and then probed with fluorescent macromolecules of three sizes. At $5\min$ , the leading edge of permeabilized epithelium featured irregular borders. At $10\min$ , full-thickness epithelial permeation was observed in some epithelial regions. Individual, nonpermeabilized cells were easily distinguished by their dark appearance. At $15\min$ , trans-epithelial penetration was observed for all three sizes of molecules. With time, the overall fluorescence intensity of the epithelium increased to approach that of the exogenous solution; however, the nuclei of cells probed with $150\mathrm{kDa}$ molecules remained dark. Fluorescence in the underlying tissues was observed as early as $5\min$ and increased steadily with time.

Fig. 2

The time course of tissue permeation following treatment with 0.5% Triton-X100. Fresh guinea pig bladder biopsies were topically treated with 0.5% Triton-X100 for $0–60\min$ and then probed fluorescent molecules of three sizes. Representative confocal images, acquired $40\mu \mathrm{m}$ below the tissue surface, are shown for the 5-, 10-, and $15\text{-}\min$ time points. The yellow/white lines indicate the apical and basal surfaces of the epithelium. The depth and uniformity of permeation increased with treatment time. Differences were observed between the cell and tissue penetration of different size probe molecules. Fluorescence signal in stroma was observed as early as $5\min$ following Triton-X100 treatment. The scale bar represents $100\mu \mathrm{m}$ .

3.2.

Depth of Mucosal Permeation

The depth of Triton-mediated permeation was assessed in a model of porcine oral mucosa ranging in thickness from $600\text{to}1000\mu \mathrm{m}$ . Epithelial punch biopsies were topically treated with different concentrations of Triton-X100 (ranging from 0 to 2.5%) for different time intervals (ranging from $0\text{to}4\mathrm{h}$ ), sliced in cross section, and probed with fluorescent macromolecules. Representative confocal fluorescent images of macromolecule permeation as a function of time can be found in Fig 3 . Figure 4 demonstrates the relationship of treatment time and Triton-X100 concentration to the depth of tissue permeation. Triton-X100 treatment was found to produce very uniform, reproducible tissue permeation with well-defined borders parallel to the treatment surface. When the leading edge of macromolecule labeling was assessed at regular time intervals, the depth of penetration was found to increase at a steady, linear rate [Fig 4a]. The smaller probe molecules consistently showed deeper penetration than the $150\mathrm{kDa}$ molecules. Macromolecule penetration to a depth of $500\mu \mathrm{m}$ was achieved in $\sim 1.5\mathrm{h}$ for molecules of $40\mathrm{kDa}$ or less, and in $\sim 4\mathrm{h}$ for molecules of $150\mathrm{kDa}$ .

Fig. 3

Permeation of the porcine oral mucosa following treatment with 1% Triton-X100. Triton-X100 was topically applied to the top surface of fresh tissue biopsies for $0–60\min$ , then the tissue was sliced, probed with fluorescent molecules of three sizes, and imaged in cross section. Representative confocal images, acquired $40\mu \mathrm{m}$ below the cut surface, are shown for the 10-, 30-, and $60\text{-}\min$ time points. Tissue permeation was characterized by a well-defined border that progressed deeper with time. The scale bar represents $100\mu \mathrm{m}$ .

Fig. 4

Permeation of normal oral mucosa as a function of time and Triton-X100 concentration. Normal porcine oral mucosa was topically treated with different concentrations of Triton-X100 for regular time intervals, sliced, and probed with fluorescent macromolecules of three sizes: $3\mathrm{kDa}$ (◼), $40\mathrm{kDa}$ (○), and $150\mathrm{kDa}$ (▲). The depth and rate of macromolecule penetration was determined from confocal fluorescence images of transverse tissue sections. (a) The relationship between the depth of permeation and duration of treatment with 1% Triton-X100. A steady, linear increase in permeation depth is observed with time. (b) Rate of tissue permeation as a function of Triton-X100 concentration. Increasing the Triton-X100 concentration increased the depth of tissue permeabilization; however, the extent of macromolecule penetration was limited by size.

Using a linear least-squares fit, the rate of tissue permeation was evaluated for a range of Triton-X100 concentrations [Fig. 4b]. Increasing the Triton-X100 concentration was found to increase the rate of tissue permeation; however, at concentrations of $>1\%$ Triton-X100, little improvement in the rate of permeation was observed. Size-dependent differences in macromolecule penetration were observed for all concentrations of Triton-X100, although there was generally no significant difference between the penetration depth of 3- and $40\mathrm{kDa}$ macromolecules. Longer probe times were evaluated to confirm that the differences in penetration depth were not due to inadequate labeling time (data not shown). Furthermore, samples in which the keratin layer was mechanically removed showed no differences in the rate of permeation when compared to intact specimens (data not shown).

3.3.

Comparison of Macromolecule Penetration in Normal and Cancer Tissue

To determine whether the depth and rate of topical permeation varies between normal and abnormal tissue, macromolecule penetration following of topical permeation was evaluated in normal oral mucosa and tumor models. Figure 5 compares the depth and rate of macromolecule penetration in two tissue models using three different labeling solutions. The maximal penetration depth following $1\mathrm{h}$ of treatment with 1% Triton-X100 was found to be $\sim 25\%$ greater in squamous carcinoma tissue than normal epithelium [Fig. 5a]. Little permeation, $<100\mu \mathrm{m}$ deep, was observed in DMSO- and PBS-treated tissues. When the rate of macromolecule penetration following Triton-X100 treatment of was evaluated as a function of size, all sizes of macromolecules consistently penetrated more rapidly in carcinoma tissue than normal epithelium [Fig. 5b]. As previously observed, the rate of macromolecule permeation was highly reproducible and linear at all time points evaluated. The largest difference in penetration rate was observed for the $150\mathrm{kDa}$ molecules, although this size of molecule consistently penetrated more slowly than smaller molecules.

Fig. 5

Comparison of macromolecule penetration in normal oral mucosa and squamous cell carcinoma. Tissues were treated with 1% Triton-X100, 10% DMSO, or PBS, sliced, and then probed with fluorescent macromolecules of three sizes. The depth and rate of macromolecule penetration was determined from confocal fluorescence images of transverse tissue sections. (a) The maximal penetration depth following a $1\text{-}\mathrm{h}$ treatment with permeation enhancers, measured using $3\mathrm{kDa}$ rhodamine-dextran. The maximal depth with Triton-X100 treatment was found to be greater for cancer tissue than normal epithelium. Little permeation was observed with DMSO- and PBS-treated controls. (b) The rate of macromolecule penetration by size with 1% Triton-X100 treatment. The rate of penetration was consistently more rapid in carcinoma tissue for all sizes of macromolecules evaluated.

3.4.

Relation of Triton-X100 Concentration to the Number of Treated Cells

The average dose of Triton-X100 per cell was estimated as function of treatment time and depth. Table 1 provides the normalized Triton-X100 concentration in terms of picomoles per cell. The calculations were based on an average cell volume, the known tissue surface area, and the measured permeation depth. In the case where only one cell layer was permeabilized ( $\sim 8.3\mu \mathrm{m}$ deep), the Triton-X100 concentration per cell was on the order of $6.2–12.3\mathrm{pmol}$ /cell. As the Triton-X100 permeated deeper, the average concentration per cell was reduced. Full thickness permeation of the bladder epithelium with a 0.5% solution yielded an average of $0.83\mathrm{pmol}$ Triton-X100 per cell. A $30–60\min$ treatment of 1.0% Triton-X100 in thicker tissues further lowered the normalized concentration to 0.28–0.57 for oral mucosa tissue and 0.21–0.42 for squamous cell carcinoma tissue.

Table 1

Triton-X100 concentration per cell as a function of treatment time and depth.

3.5.

Viability Assessment of Tissue Phantoms

Tissue phantoms were assayed for cell viability following treatment with different concentrations of Triton-X100. Figure 6 shows representative images of phantoms, containing five million cells each, treated with 0–5% Triton-X100 for $1\mathrm{h}$ and then coincubated with MTT reagent. The normalized Triton-X100 concentration per cell was determined from the number of cells incorporated into each phantom. Phantoms treated with 1% Triton-X100 or less ( $⩽3.3\mathrm{pmol}$ /cell) demonstrated an intense blue color indicative of MTT reduction by viable cells. When imaged in cross section, these phantoms showed intense staining throughout. In contrast, phantoms treated with concentrations above 1% ( $>3.3\mathrm{pmol}$ /cell) demonstrated an overall loss in color intensity as well as a lack of staining at the phantom periphery, suggestive of cell death at these higher Triton-X100 concentrations.

Fig. 6

Tissue phantoms assayed for viability after treatment with different concentrations of Triton-X100. Phantoms containing $5\times {10}^6$ cells each were treated with $1\mathrm{mL}$ Triton-X100 for $1\mathrm{h}$ , returned to media for $4\mathrm{h}$ , and then coincubated with a MTT reagent for $20\min$ . Blue color deposition is indicative of viable cells. Intense staining was observed in phantoms treated with 1% Triton-X100 or less ( $⩽3.3\mathrm{pmol}$ /cell). Phantoms treated with Triton-X100 concentrations above 1% ( $>3.3\mathrm{pmol}$ /cell) demonstrated an overall loss in color intensity and staining at the periphery.

3.6.

EGFR Targeting in Cancer Tissue

To determine whether Triton-X100 is useful for the delivery of molecular-specific optical contrast agents, subcutaneous tumors were labeled for EGFR in the presence of topical permeation enhancers. Freshly excised tumors were topically treated for $2\mathrm{h}$ with antiEGFR-647 diluted in 1% Triton-X100, 10% DMSO, or PBS. The labeled tissues were sliced in cross section, counterstained with acriflavine-HCL, and imaged perpendicular to the labeling surface. Figure 7 shows representative confocal fluorescence images of EGFR labeling in the presence of different topical permeation enhancers. The surface of the tissue where the contrast agent was applied appears at the top of each panel. EGFR labeling is denoted in red and acriflavine-HCL labeling in green. Treatment with 1% Triton-X100 facilitated the uniform delivery of antiEGFR-647 to a depth of $500–600\mu \mathrm{m}$ . Washing of the tissue after contrast agent delivery yielded a honeycomb pattern of labeling consistent with that observed in vitro. EGFR labeling was reproducibly observed around all squamous cells to the full depth of contrast agent delivery. The intensity of labeling was similar to that observed in cell monolayers labeled in the presence and absence of Triton-X100 (data not shown). In contrast, EGFR labeling in DMSO- and PBS-treated tissues was limited to the superficial layers of cells ( $\sim 20–60\mu \mathrm{m}$ deep).

Fig. 7

EGFR targeting in the presence of topical permeation enhancers. Subcutaneously generated 1483 tumors were topically labeled for $2\mathrm{h}$ with antiEGFR-647 (red) diluted in 1% Triton-X100, 10% DMSO, or PBS. The tissues were sliced and imaged perpendicular to the labeling surface, shown at the top. Tissues were counterstained with acriflavine-HCL (green) to highlight the tissue morphology. In Triton-X100 treated tissues, uniform EGFR labeling was observed around all squamous cells to a depth of $500–600\mu \mathrm{m}$ . EGFR labeling in the DMSO- and PBS-treated controls was limited to the superficial cell layers. The scale bar represents $100\mu \mathrm{m}$ .

3.7.

Multitarget Labeling in Tumors

To demonstrate the feasibility of delivering multiple optical contrast agents to different cell compartments, three different targeting moieties were topically applied to tumors in the presence of Triton-X100. The specificity and bioactivity of each targeting moiety was established in vitro prior to tissue labeling. No differences in targeting were observed with Triton-X100 co-treatment of live and fixed cells (data not shown). Figure 8 shows representative confocal images of freshly excised tumor tissue simultaneously labeled for EGFR, Cyclin D1, and nucleic acid content in the presence of 1% Triton-X100. The tissues were topically labeled, washed, sliced, and imaged perpendicular to the labeling surface. AntiEGFR-647 labeling (red) appeared in a characteristic honeycomb pattern. AntiCyclin D1-488 labeling (green) was diffusely spread through the cytoplasm of cancer cells. Propidium iodide labeling (orange) appeared in the nuclei of permeabilized cells. EGFR, Cyclin D1, and propidium iodide labeling always appeared together (overlay), suggesting that the labeling is limited to cells permeabilized by the Triton-X100 treatment. Little to no antibody labeling was observed when isotype controls and propidium iodide were administered in the presence of Triton-X100, suggesting that unbound contrast agents were sufficiently removed with the washing steps. No propidium iodide labeling was present in tissues treated with PBS instead of Triton-X100 (data not shown).

Fig. 8

Cotargeting of clinically relevant cancer biomarkers. Subcutaneously generated 1483 tumors were topically labeled for $1\mathrm{h}$ with three molecular-specific contrast agents diluted in 1.0% Triton-X100. The contrast agents were targeted to EGFR on the cell membrane (antiEGFR-647, red), Cyclin D1 in the cytoplasm (antiCyclin D1, green), and nucleic acids in the nucleus (propidium iodide, orange). After labeling, the tissues were washed, sliced, and imaged perpendicular to the labeling surface. The contrast agents were found to localize to the appropriate cell compartments. EGFR, Cyclin D1, and nucleic acid labeling always appeared together in treated tissue (overlay). The observed pattern of labeling was highly uniform and reproducible from sample to sample. Little to no labeling was observed in IgG-treated controls with Triton-X100 cotreatment, suggesting that the washing of the tissue was sufficient to remove unbound contrast agents. The scale bar represents $20\mu \mathrm{m}$ .