Targeted Fluorogenic Cyanine Carbamates Enable In Vivo Analysis of Antibody–Drug Conjugate Linker Chemistry

Abstract

Antibody–drug conjugates (ADCs) are a rapidly emerging therapeutic platform. The chemical linker between the antibody and the drug payload plays an essential role in the efficacy and tolerability of these agents. New methods that quantitatively assess the cleavage efficiency in complex tissue settings could provide valuable insights into the ADC design process. Here we report the development of a near-infrared (NIR) optical imaging approach that measures the site and extent of linker cleavage in mouse models. This approach is enabled by a superior variant of our recently devised cyanine carbamate (CyBam) platform. We identify a novel tertiary amine-containing norcyanine, the product of CyBam cleavage, that exhibits a dramatically increased cellular signal due to an improved cellular permeability and lysosomal accumulation. The resulting cyanine lysosome-targeting carbamates (CyLBams) are ~50× brighter in cells, and we find this strategy is essential for high-contrast in vivo targeted imaging. Finally, we compare a panel of several common ADC linkers across two antibodies and tumor models. These studies indicate that cathepsin-cleavable linkers provide dramatically higher tumor activation relative to hindered or nonhindered disulfides, an observation that is only apparent with in vivo imaging. This strategy enables quantitative comparisons of cleavable linker chemistries in complex tissue settings with implications across the drug delivery landscape.

Graphical Abstract

INTRODUCTION

Antibody–drug conjugates (ADCs) combine the specificity of monoclonal antibodies (mAbs) and the potency of small-molecule therapeutics. To date, 10 ADCs have received FDA approval and many others (>60) are in clinical trials.^1–3 ADC activity generally requires lysosomal processing of a linker domain to release the active payload. Consequently, the linker component should be stable in circulation but selectively cleaved following target binding and internalization - a significant chemical challenge.^4–8

To assess the state-of-the-art and to guide future linker discovery efforts, a way to quantitatively compare ADC linker chemistry in vivo would be of significant utility. ADCs are conventionally assessed by examining tumoricidal activity and toxicity profiling. While these methods are important benchmarks, they provide only indirect insights into the site and mechanism of drug release.^9–11 Enzyme-linked immunosorbent assays (ELISAs) are also broadly employed but only determine the blood-pool concentration and biodistribution of the antibody component.^12,13 Radiolabeling methods can provide important insights regarding mAb localization but are costly and do not directly report on the linker cleavage step.^14–17

Optical imaging has the potential to provide critical insights to the ADC design and optimization process. Prior efforts using stimuli-responsive fluorophores with conventional visible wavelengths have quantified payload processing and internalization kinetics in cellular imaging experiments.^18–20 However, these probes are not suitable for applications in tissue due to the poor penetration depth of wavelengths in the visible region. Complementing these efforts, approaches using always-ON near-infrared (NIR) probes have provided insights into tumor and off-target uptake.^21–25 Fluorogenic turn-ON probes that use NIR wavelengths (~700–900 nm) have significant potential to provide insight into the dynamics and localization of biological phenomena in complex tissue settings. We recently developed the first class of NIR fluorogenic probes based on the heptamethine cyanine core.²⁶ These fluorogenic cyanine carbamates (CyBams, Figure 1A) exhibit exceptional turn-ON ratios, and untargeted variants enabled in vivo imaging in a metastatic tumor model. In these studies, we hypothesized that mAb-conjugated variants could provide a real-time quantitative means to determine the site and extent of ADC linker cleavage in complex model organisms.

Figure 1.

(A) Evolution and (B) overview of CyLBam chemistry.

Here we detail the development of the first activatable mAb-targeted probes with absorbance and emission maxima beyond 700 nm. To obtain a sufficient signal for in vivo imaging, we optimized the cellular uptake and retention of the released fluorescent product of linker cleavage, the pH-sensitive norcyanine. These efforts reveal that the installation of a basic amine into the probe dramatically improves the cellular photon output, likely due to enhanced lysosomal uptake and retention. Converting the optimized norcyanine into the corresponding cyanine lysosome-targeting carbamate (CyLBam, Figure 1A,B) dramatically improves the in vitro signal and is essential for high-contrast in vivo optical imaging of mAb conjugates. Finally, we test a panel of commonly used ADC linkers in two tumor models. We find that cathepsin-cleavable linkers outperform reductively cleaved disulfide linkers with dramatic differences in tumor uptake, a distinction that is only apparent with in vivo imaging. Broadly, this approach provides a general means to assess cleavable linker efficiency and specificity, with applications from the cellular to organismal scales.

RESULTS AND DISCUSSION

Optimization of the Norcyanine Signal.

Prior to pursuing in vivo imaging of mAb conjugates, we set out to improve the cellular signal of the product of CyBam activation, the pH-sensitive norcyanine. While the previously reported sulfonated norcyanine, Sulfo-NorCy7 (Figure 2A), can provide high-contrast imaging, relatively high concentrations (20–40 μM) are required for a sufficient in vitro signal.^26–29 We anticipated that this requirement might be problematic for mAb-targeted imaging because probe concentrations are intrinsically limited by antigen levels. Of note, microscopy studies indicate that the cellular signal of Sulfo-NorCy7 is predominately from the lysosome. The cellular signal is likely only due to the lysosomal fraction, given the observed pK_a between 4 and 5.²⁶ We therefore assume poor lysosomal accumulation is responsible for the modest cellular fluorescence. We hypothesized that by increasing the cellular permeability and perhaps introducing a lysosomal-targeting element, we could improve the fluorescent output.

Figure 2.

(A) Structures of norcyanines Sulfo-, H-, OMe-, Pip-, N-Me-Pip-, and N-Me-NorCy7 (X, left, blue; R, right, red). (B) Absorbance and fluorescence spectra of N-Me-NorCy7 (5 μM) in PBS pH 7.20 and acetate buffer pH 4.25 (ex. 690 nm). (C) Determination of pK_a (10 μM) in buffers ranging from pH 3.5 to pH 7.4. (D) Flow cytometry quantification of the in vitro uptake of norcyanine heptamethine cyanines (5 μM) in MDA-MB-468 after 1 h of incubation. The geometric mean fluorescent intensity (± SD) of the fluorescent signal in the cells is shown (n = 4 independent experiments; ~10 000 cells counted). (E) Confocal fluorescent images (63X) of N-Me-NorCy7 and Sulfo-NorCy7 (10 μM) after 6 h of incubation with MDA-MB-468. The fluorescent signals from the probe and nucleus (Hoechst) are pseudocolored red and blue, respectively. Scale bar is 10 μm.

To approach this problem, we designed and synthesized a small panel of chemically diverse norcyanines. These compounds contained a C4′-phenyl-substituent appended to the polymethine chromophore, which we and others have found improves the synthetic accessibility, photostability, and serum stability.^30–35 To test the question of cell permeability alone, we included two hydrophobic derivatives, OMe-NorCy7 and H-Nor-Cy7 (Figure 2A). Commonly used lysotracker probes contain tertiary amines, suggesting this functional group can promote lysosomal targeting.^36–38 To test if this approach might apply to norcyanines, we prepared indolenine-substituted sulfonamide derivatives modified with both the hydrophobic piperidine (Pip-NorCy7) and the tertiary amine-containing N-Me piperazine (N-Me-Pip-NorCy7) substituents. We also prepared another “lysotracker-like” derivative, N-Me-NorCy7, in which the central ring system contains an N-methyl substituent. These novel norcyanines were synthesized in three or four steps, as described in the Supporting Information.

With access to this panel of probes, we investigated their photophysical properties and cellular uptake. All six probes exhibit absorbance and emission maxima above 700 nm, similar extinction coefficients and quantum yields, and pK_a values between 4.4 and 5.8 (Figure 2B and C for N-Me-NorCy7, Table 1, and Figure S1). We first established that the NorCy7 series had a minimal toxicity in MDA-MB-468 and MCF-7 cells (Figure S2). Next, we quantified signal in MDA-MB-468 cells at 1, 3, and 6 h time points with flow cytometry using the APC-Cy7 channel. At all three time points, the uptake of the six norcyanines follows the rank order N-Me-NorCy7 > N-Me-Pip-NorCy7 ≫ Pip-NorCy7, OMe-NorCy7, H-Nor-Cy7 ≫ Sulfo-NorCy7 (Figures 2D, S3, and S31–33). To our delight, N-Me-NorCy7 and N-Me-Pip-NorCy7 exhibited 187- and 72-fold higher uptake, respectively, in MDA-MB-468 cells after only 1 h incubation compared to the previously used Sulfo-NorCy7. The fluorescent signal remained constant at the 1, 3, and 6 h time points, suggesting efficient uptake and high retention of tertiary amine-substituted norcyanines (N-Me-NorCy7 and N-Me-Pip-NorCy7). In contrast, the signal from Sulfo-NorCy7 increased steadily overtime and was around threefold higher after 6 h relative to the level at 1 h (Figure S3). Using confocal microscopy, we confirmed that the subcellular localization of N-Me-NorCy7 is lysosomal, with a dramatically higher signal than that of Sulfo-NorCy7 (Figures 2E, S4, and S5). The relatively modest signal of the three hydrophobic derivatives, namely Pip-NorCy7, OMe-NorCy7, and H-NorCy7, indicates that the lysotracker-mimicking tertiary amine is essential for fluorescent signal enhancement. Overall, these data indicate that the incorporation of a single basic amine dramatically enhances the cellular uptake and lysosomal localization of norcyanines. The enhanced cellular signal and synthetic accessibility led us to choose N-Me-NorCy7 to compare to the previously described Sulfo-NorCy7 in the generation of activable mAb conjugates.

Table 1.

Summary of Photophysical Properties of the NorCy7 Series of Compounds

	λmax,abs/λmax,em (nm)	εmax (M−1 cm−1)	pKa	ΦFd(%)
Sulfo-NorCy7a	755/775	73 500	5.2	8.9
H-NorCy7b	750/770	63 900	5.8	9.1
OMe-NorCy7b	775/790	56 800	5.5	1.6
Pip-NorCy7c	760/780	17 200	4.4	8.1
N-Me-Pip-NorCy7a	760/775	45 000	4.5	10
N-Me-NorCy7a	730/750	57 600	4.6	12

^aAcetate buffer pH 4.5.

^bMeOH/acetate buffer pH 4.5 (2:1).

^cMeOH/acetate buffer pH 3.75 (2:1).

^dMeOH + 0.1% formic acid for all compounds.

mAb-Targeted Imaging.

We then assessed the impact of norcyanine modification on the signal of mAb conjugates. N-Me-NorCy7 was converted to the cyanine lysosome-targeting carbamate (CyLBam) variant by attaching a cathepsin-cleavable dipeptide (valine–citrulline, VC) and a noncleavable (NC) glutaric anhydride linker substituted with a lysine-reactive NHS ester (see the SI for synthetic details). As a comparison, Sulfo-NorCy7 was converted to the corresponding CyBam molecules. The four probes were conjugated to the FDA-approved anti-epidermal growth factor receptor (EGFR) antibody panitumumab (Pan) at pH 7.4 (degree of labeling (DOL) ~ 4; Tables S1 and S28). The resulting conjugates were purified by both spin column and dialysis to ensure that no free small molecule remained. Of note, prior work from our lab and others has shown that similar conjugates maintain the in vitro and in vivo targeting properties of the parent mAb.^{22,5,21,39–41}

We first compared the series of Pan probes in cellular assays. The four conjugates, namely Pan-VC-CyLBam, Pan-NC-CyLBam, Pan-VC-CyBam, and Pan-NC-CyBam, (Figure 3A) were tested in MDA-MB-468 (EGFR+) and MCF-7 (EGFR−) cells.^42,43 The fluorescent signal emitted from Pan-VC-CyLBam was 50-fold higher than that of Pan-VC-CyBam after incubation with MDA-MB-468 cells for 24 h (Figures 3C and S34–37). Validating the role of receptor-mediated uptake, the fluorescent signal of both probes was lower in MCF-7 cells (by 7.5-fold for the CyLBam and 4.5-fold for the CyBam (Figure 3D)). As expected, the noncleavable probes Pan-NC-CyLBam and Pan-NC-CyBam did not exhibit any significant fluorescent signal in either cell line. Confocal microscopy confirmed the trends observed by flow cytometry as well as the lysosomal uptake of the released norcyanine (Figures 3B and S6–8).

Figure 3.

(A) Structures of CyLBam and CyBam probes and VC (val-cit) and NC (noncleavable) linkers. (B) Confocal images (63×) of Pan-VC-CyLBam, Pan-NC-CyLBam, Pan-VC-CyBam, and Pan-NC-CyBam in MDA-MB-468 (EGFR+) after 24 h of incubation. Fluorescent signals from the probe and nucleus (Hoechst) are shown in red and blue, respectively. Scale bar is 10 μm. (C) Flow Cytometry quantification of in vitro uptake in MDA-MB-468 (EGFR+) and MCF-7 (EGFR−) after 24 h of incubation. All antibody conjugates were labeled at DOL 4. The geometric mean fluorescent intensity (±SD) of fluorescent signal in the cells is shown (n = 4 independent experiments; ~10 000 cells counted). (D) Imaging of probes (200 μg; DOL 4) intravenously injected in a xenograft model of female athymic nude mice implanted with MDA-MB-468 tumors (n = 3). Tumor-to-background (TBR) ratios, the fluorescent signal of the tumor relative to an equal area in the neck, are shown at different time intervals (4, 48, and 168 h). (E) IVIS fluorescent images 48 h post-injection. Tumors are highlighted in dotted red circles. Data points are displayed as the mean ± SD. The p-values were evaluated by the Student’s t-test (***p ≤ 0.001, ****p ≤ 0.0001).

We then set out to test if the improved in vitro signal of the CyLBam conjugate would translate to high-contrast animal imaging. The conjugates (200 μg) were injected intravenously into female athymic nude mice with MDA-MB-468 xenograft tumors (25–35 mm³). The mice were imaged at 4, 24, 48, 72, and 168 h post-injection using an in vivo imaging system (IVIS) (Figures 3E and S9–14). After 24 h, a strong fluorescent signal was observed in tumors injected with Pan-VC-CyLBam, and the signal persisted over 168 h, with tumor-to-background ratios (TBRs) between 4 and 5. In contrast, the Pan-VC-CyBam group had lower TBRs (1–2) at all time points, similar to the two noncleavable probes Pan-NC-CyLBam and Pan-NC-CyBam. Significant liver signals were observed with the two probes containing the protease-cleavable linker (Pan-VC-CyLBam and Pan-VC-CyBam) at early time points (4 and 24 h), and the fluorescent signal decreased afterward (Figures S15 and S16). The selective hepatic signal of Pan-VC-CyBam and its time-dependent disappearance suggests that the released Sulfo-NorCy7 is rapidly cleared through hepatobiliary pathways. Thus, the reduced tumor signal observed with this probe likely reflects a lower tumor retention (compared to N-Me-NorCy7) and subsequent clearance. Overall, these results suggest that the improved cellular retention of N-Me-NorCy7, first observed in vitro, extends to an in vivo setting. Critically, these observations set the stage for us to evaluate a series of ADC linkers.

Quantitative Comparison of ADC Linkers.

The two most extensively studied classes of ADC linkers are disulfides, which are reductively cleaved by biological thiols, and peptide linkers, which are cleaved by lysosomal cathepsins. Significant efforts to apply disulfide linkers led to the approval and clinical testing of several agents.⁴⁴ These include the first approved ADC, the acute myeloid leukemia therapy gemtuzumab ozogamicin (Mylotarg), which contains a hindered disulfide.^45–47 However, some reports suggest that certain disulfides can be cleaved in circulation, leading to premature payload release and nonspecific uptake.^47–50 Concurrent with these efforts, cathepsin-cleavable peptide linkers have been incorporated in four approved ADCs, including the Hodgkins lymphoma drug brentuximab vedotin (Adcetris). While these two classes of linkers have been investigated extensively, we are not aware of any reports directly comparing the relative cleavage of payloads in a solid tumor setting. To do this, we designed a small panel of probes comparing two commonly used cathepsin-cleavable peptide linkers, namely valine–citrulline Pan-VC-CyLBam (used above) and alanine–alanine, Pan-AA-CyLBam, and two disulfides, one hindered gem-dimethyl-substituted variant Pan-S,S-Me₂-CyLBam and one primary disulfide Pan-S,S-CyLBam (Figure 4A and Table S2).^9,51 As a control, the noncleavable probe described above Pan-NC-CyLBam, was also employed.

Figure 4.

(A) Structures of Pan-AA-CyLBam, Pan-S,S-CyLBam, and Pan-S,S-Me₂-CyLBam. (B) Quantification of the in vitro uptake of the five probes (50 μg; DOL 4) in MDA-MB-468 (EGFR+) and MCF-7 (EGFR−) after 24 h of incubation. The geometric mean fluorescent intensity (±SD) of the fluorescent signal in the cells is shown (n = 4 independent experiments; ~10 000 cells counted). (C) Fluorescent images following the injection of probes (100 μg; DOL 4) in female athymic nude mice (n = 5) with MDA-MB-468 tumors at the 48 h time point. Tumors are highlighted in dotted white circles. (D) Quantification of TBR 48 h post-injection. (E) Quantification of the fluorescent signal of Pan-AA-CyLBam from the liver and the tumor. The tumor-to-liver ratio (TLR) is depicted in the insert. Data points are displayed as mean ± SD. The p-values were evaluated by the Student’s t-test (**p ≤ 0.01, ***p ≤ 0.001).

We first examined the signal of the panel of conjugates in cells expressing variable levels of EGFR. The fluorescent signal from linker cleavage was quantified using flow cytometry at 6 and 24 h post-incubation (Figures 4B and S38–41). At the 6 h time point, Pan-S,S-CyLBam exhibited the highest fluorescent signal in EGFR+ MDA-MB-468 cells. The other three cleavable probes had only modest fluorescent signals after 6 h, but the signal increased substantially after 24 h. Pan-VC-CyLBam had the highest fluorescent signal among the cleavable probes after 24 h, indicating the most extensive linker cleavage. All five conjugates exhibit low levels of the probe signal at 6 and 24 h in the EGFR− cell lines (MCF-7 and MDA-MB-231, Figures 4B, S17, S42, and S43), albeit with a slightly higher signal for the Pan-S,S-CyLBam probe. Significantly, if these data were used in isolation, it would be difficult to differentiate these linkers and assess the optimal agent for further study.

With the goal of investigating differences between these linkers in an organismal context, we then harnessed the NIR spectroscopic properties of these probes for in vivo imaging. The full series of probes (100 μg dose, half the dose of the studies above), were administered to female athymic nude mice that were implanted with MDA-MB-468 tumors and imaged at regular intervals up to 48 h (Figure 4C). With the disulfide probes Pan-S,S-Me₂-CyLBam and Pan-S,S-CyLBam, we did not observe significant tumor signal, and the TBR was similar to that of the noncleavable Pan-NC-CyLBam control. In contrast, strong tumor signals were observed from the cathepsin-sensitive probes Pan-VC-CyLBam and Pan-AA-CyLBam (Figures 4C and D and S19–21), with the TBR reaching as high as 4.5 after 48 h. Therefore, these studies indicate that cathepsin-cleavable linkers provide dramatically higher tumor activation relative to hindered or nonhindered disulfides. Critically, as only modest differences were observed using the cellular methods described above, this key distinction is only apparent with in vivo imaging.

In addition to tumor uptake, these data also provide significant insight into off-target cleavage pathways. As observed above, conjugates with cleavable linkers exhibit significant liver signals at early time points (4 and 24 h, Figure S18). While likely due to a combination of liver uptake/cleavage and probe clearance through hepatobiliary pathways, this approach can provide quantitative insight into off-target cleavage. With Pan-AA-CyLBam as an example, the liver signal decreases overtime but increases in the tumor (Figure 4E), which leads to an increase in the tumor-to-liver ratio (TLR) over time (Figure 4E, insert). These results are in line with prior observations that suggested significant liver linker cleavage when related cathepsin linkers were used.^52,53 We also corroborated the in vivo imaging results with an ex vivo assessment, which was performed 48 h post injection. As expected, the TMR (tumor-to-muscle ratio) revealed a trend identical to that observed in vivo (Figures S22 and S23).

Finally, to test the generality of our approach, CyLBams were tested against an alternative cancer target. The cell surface protein CD276, also known as B7 homologue H3 (B7H3), is overexpressed in both tumor cells and the tumor vasculature of multiple solid tumor types.^54–57 Antibodies against this promising target have been used to develop ADCs armed with either pyrrolobenzodiazepine- (PBD) or auristatin-derived (MMAE) payloads. These ADCs, which are constructed using cathepsin-cleavable linkers, exhibit potent antitumor activity.^58,59 We chose to compare VC and S,S-Me₂ linkers and prepared the CD276 targeting antibody m276-SL reported previously, the noncleavable NC control m276-SL conjugates, and nonbinding mAb IgG controls (DOL 4, Table S3 and Figures S29 and S30). In initial in vitro tests using a CD276+ JIMT-1 triple-negative breast cancer cell line, both cleavable m276-SL conjugates showed high levels of cleavage, and little activation was shown using the control nonbinding IgG antibody (Figures 5A, S24, S44, and S45). We then carried out a study comparing protease-cleavable m276-SL-VC-CyLBam and disulfide-cleavable m276-SL-S,S-Me₂-CyLBam conjugates in JIMT-1 tumors (200–250 mm³) grown orthotopically in the mammary fat pad. As observed in the studies above, only the protease-cleavable m276-SL-VC-CyLBam conjugate exhibited a TBR greater than 5 (5.2 after 48 h), while the disulfide m276-SL-S,S-Me₂-CyLBam probe exhibited a much lower tumor signal (Figures 5B and C and S25–27). These initial studies in this model serve to validate both the generality of our CyBam imaging strategy and the observation that protease-cleavable linkers outperform disulfide linkers in solid tumor models.

Figure 5.

(A) Quantification of the cellular uptake of m276-SL-S,SMe₂-CyLBam, m276-SL-VC-CyLBam, and m276-SL-NC-CyLBam (80 μg; DOL 4) in JIMT-1 cells following 24 h of incubation. The geometric mean fluorescent intensity (±SD) of the fluorescent signal in the cells is shown (n = 4 independent experiments; ~10 000 cells counted). (B) TBRs at different time intervals after the intravenous injection of m276-SL-S,SMe₂-CyLBam and m276-SL-VC-CyLBam (100 μg; DOL ~4) in female athymic nude mice bearing orthotopic JIMT-1 tumors (n = 5). (C) Fluorescent images 72 h post-injection. Tumors are highlighted in dotted red circles.

CONCLUSIONS

The mode of action of ADCs is complex. Optimal efficacy depends on target binding, internalization, and finally lysosomal catabolism to initiate the payload release. While the linker domain plays a central role in determining the efficacy and selectivity of ADCs, it is challenging to directly assess the site and extent of linker cleavage using conventional methods.^6–8 This is a critical issue because off-target cleavage of ADCs contributes significantly to toxicity.⁶⁰ The optical imaging approach detailed here provides a general means to analyze linker chemistry across cellular, tissue, and body-wide scales. In this work, we identify probes suitable for mAb-targeted activatable imaging. While our previously reported sulfonated CyBams were not suitable for this application, the installation of a tertiary amine into the norcyanine framework dramatically improves cellular and in vivo signals, likely due to enhanced lysosomal uptake and retention. It is likely that these “lysotracker-like” norcyanines and the resulting CyLBam probes will have applications in other settings, including their use as activity-based sensing agents.^61,62

In applying this approach for ADC applications, we assume that the fluorescent probe can serve as a useful surrogate for the payload component. Here we have shown that quantitative comparisons of linker chemistries can be obtained by varying the linker component. In addition to linker chemistry, the properties of the payload molecule (e.g., charge and hydrophobicity) are also important aspects of tumor targeting.^63–66 Critically, while the norcyanine probe is a hydrophobic sp²-rich small molecule, so too are many ADC payloads.^67–70 Further variations of the cyanine component could provide additional insights into the role of payload properties in tumor and normal tissue distribution, something our group has examined previously in the context of always-ON fluorophore conjugates.^41,71–74

Going forward, we anticipate that CyLBam imaging could provide insights that directly inform the development of ADCs. These studies have suggested significant differences between protease-cleavable dipeptides and disulfides in terms of tumor cleavage, which is likely due to instability of these disulfides in circulation. Prior studies looking at various cysteine mutants have found the site of mAb labeling can have dramatic effects on the disulfide stability.⁷⁵ This approach is well positioned to investigate these and related homogeneous labeling strategies. We also anticipate this strategy can be readily extended to analyze the impact of mAb engineering strategies.^76–78 As altering the molecular weight and composition of the protein component can dramatically alter the clearance pathway, it is likely that linker cleavage dynamics will be impacted. We also anticipate applications in investigating novel linker cleavage strategies, such as those that rely on extracellular proteases. Finally, we note that this strategy can readily be translated to other targeted drug delivery strategies that rely on cleavable linkers. Efforts toward these goals, and to translate lessons from these imaging efforts into novel therapeutics, are underway.

ABBREVIATIONS

NIR

near-infrared

ADC

antibody–drug conjugate

mAb

monoclonal antibody

CyBam

cyanine carbamate

CyLBam

cyanine lysosomal carbamate

ELISA

enzyme-linked immunosorbent assay

PBS

phosphate buffer saline

B7H3

B7 homologue H3

IVIS

in vivo imaging system

CD276

cluster of differentiation 276

EGFR

epidermal growth factor

abs

absorbance

ex

excitation

em

emission

NA

numerical aperture

VC

valine–citrulline

AA

alanine–alanine

NC

noncleavable

Pan

panitumumab

NHS

N-hydroxy succinimide

DOL

degree of labeling

TBR

tumor-to-background ratio

TLR

liver-to-liver ratio

TMR

tumor-to-muscle ratio

IgG

immunoglobulin G

MeOH

methanol

SD

standard deviation