Introduction

QDs are fluorescent semiconductor nanoparticles with many unique optical properties suitable for multiplexed in vitro and in vivo imaging^{1, 2, 3}. Numerous in vitro and cell-based applications have been discovered for QDs^{4, 5}. In the near-infrared (700–900 nm) region, the absorbance spectra for most biomolecules reach minimum, which provides a clear window for in vivo optical imaging⁶. A vast number of literature is available where nontargeted QDs were used for cell trafficking^{7, 8}, vasculature imaging^{9, 10}, sentinel lymph node mapping^{11, 12} and neural imaging¹³. To make QDs more useful for in vivo imaging and other biomedical applications, QDs need to be effectively, specifically and reliably directed to a specific organ or disease site without alteration. Specific targeting can be achieved by attaching targeting molecules to the QD surface. Peptides, peptidomimetics and small molecules are suitable targeting ligands, as large numbers of these molecules can be linked to the surface of a single QD and such conjugates may exhibit strong receptor binding affinity and desirable targeting efficacy due to the polyvalency effect¹⁴.

Angiogenesis, the formation of new blood vessels from the preexisting vasculature, is essential for tumor growth and progression^{15, 16, 17}. Integrin _v3, a cell adhesion molecule, is significantly upregulated in invasive tumor cells of many cancer types and most tumor vasculature but not in quiescent endothelium and normal tissues^{18, 19}. The fact that integrin _v3 is overexpressed on both tumor vasculature and tumor cells makes it a prime target for in vivo targeted imaging using QD-based probes, as extravasation is not required to observe tumor signal. We recently reported the in vivo targeted imaging of tumor vasculature using peptide-conjugated QDs²⁰. In this study, RGD (potent integrin _v3 antagonist)-containing peptides were conjugated to QD705 (emission maximum at 705 nm) and QD705-RGD exhibited high-affinity integrin _v3 specific binding in cell culture, ex vivo, and in living mice bearing subcutaneous integrin _v3-positive U87MG human glioblastoma tumors.

The protocol described here provides a step-by-step procedure for the preparation of this QD705-RGD conjugate and its subsequent use for cell staining and in vivo targeted imaging. The chemistry of preparing this conjugate can be employed to attach other thiol-containing ligands to any amino functionalized QDs (Fig. 1). The QD705 used in this study contains a polyethylene glycol (PEG; MW 2 kDa) spacer covalently attached to the QD surface, resulting in improved stability in high-salt buffers and reduced nonspecific binding. We found that carboxyl-modified QDs that do not contain PEG spacers may precipitate from high-salt buffers. Moreover, the significantly more nonspecific binding to cells compared to the amino functionalized QDs also makes the experimental findings harder to interpret. As both QD705 and the RGD peptide contain amino groups, conjugation through a heterobifunctional linker (e.g., 4-maleimidobutyric acid N-succinimidyl ester) after converting one of the amino group to a thiol group is the most convenient and widely used approach²¹. Alternatively, reacting excess amount of a homobifunctional linker (e.g., one with two N-succinimidyl esters) with either QD705 or the RGD peptide, which, upon purification, is conjugated to the other agent, may also be employed. However, this approach was not widely used in the literature, and it is also difficult to control the reaction in certain cases.

Figure 1: Synthesis of QD705-RGD.

equiv., equivalent; PEG, polyethylene glycol (MW 2 kDa).

Full size image (72 KB)

The major limitation of this protocol does not lie in the chemistry itself but in choosing the appropriate target. QD705, as well as QD705-RGD, has rather short circulation half-life and very rapid uptake in the reticuloendothelial system. Thus, there is not enough circulation time to allow for efficient extravasation of QD705-RGD to target the tumor cells. The expression level of integrin _v3 is very high on the U87MG tumor vasculature^{22, 23}. Therefore, even with almost exclusive tumor vasculature targeting, we were able to observe good tumor-to-background contrast. With future development of small, biocompatible, long-circulating QDs, it may become possible to target both the tumor vasculature and the tumor cells.

For in vitro testing of QD705-RGD, cells with high and low integrin _v3 expression, for example, U87MG human glioblastoma and MCF-7 human breast cancer cells^{22, 24}, can serve as positive and negative controls. For in vivo imaging, tumors with low integrin _v3 expression, as well as QD705 without RGD conjugation, can serve as valid negative controls. Both in vitro and in vivo receptor blocking experiments with unconjugated RGD peptide can also be performed to further confirm the integrin _v3 specificity of QD705-RGD. To obtain statistically meaningful results, at least three samples or animals per group are needed.

QD-based NIRF imaging in small animals cannot be directly scaled up to in vivo imaging in patients due to the limited optical signal penetration depth. In clinical settings, fluorescence imaging is relevant for tissues close to the surface of the skin and tissues accessible by endoscopy and intraoperative visualization. The major roadblocks for clinical translation of QD-based agents are inefficient delivery, potential toxicity and lack of quantification¹. However, with the development of smaller^{12, 25}, less-toxic²⁶, self-illuminating^{27, 28}, multifunctional QDs^{29, 30} and further improvement of the conjugation strategy, it is expected that QD-based probes may achieve optimal tumor-targeting efficacy with acceptable toxicity profile for clinical translation in the near future.

Materials

Reagents

• QDs (Qdot 705 ITK amino (PEG) QDs; Invitrogen, cat. no. Q21561MP)
• 4-Maleimidobutyric acid N-succinimidyl ester (Sigma, cat. no. M7642)
• 10 mM sodium borate buffer, pH 8.5 (Sigma, cat. no. 71997)
• PBS buffer (Invitrogen, cat. no. 20012-027)
• RGD peptide (c(RGDyK); Peptides International)
• N-Succinimidyl S-acetylthioacetate (Pierce Biotechnology Inc., cat. no. 26102)
• DMSO (ACROS, cat. no. 61097-1000)
• High-performance liquid chromatography (HPLC)-grade acetonitrile (Fisher Scientific, cat. no. A998-4)
• Trifluoroacetic acid (Fisher Scientific, cat. no. AC139721000)
• Hydroxylamine hydrochloride (Sigma, cat. no. 255580)
• Tris(2-carboxyethyl)phosphine hydrochloride (TCEP.HCl; Pierce Biotechnology Inc., cat. no. 20490)
• Sodium hydroxide (Sigma, cat. no. 367176)
• Hydrochloric acid (Sigma, cat. no. H3162)
• Thiolated RGD peptide (see Box 1 and Fig. 2)

  Figure 2: Converting the amino group in the RGD peptide into a thiol.

  

  Full size image (40 KB)
  
  
• U87MG human glioblastoma cells (ATCC, cat. no. HTB-14; see REAGENT SETUP)
• MCF-7 human breast cancer cells (ATCC, cat. no. HTB-22; see REAGENT SETUP)
• Low-glucose DMEM (GIBCO, cat. no. 11885-084)
• Minimum essential medium (GIBCO, cat. no. 11095-080)
• Fetal bovine serum (FBS; GIBCO, cat. no. 10437-028)
• Cell-binding buffer (20 mM Tris, 150 mM NaCl, 2 mM CaCl₂, 1 mM MnCl₂, 1 mM MgCl₂, 0.1% (wt/vol) bovine serum albumin; pH 7.4)
• ¹²⁵I-Echistatin (GE Healthcare, cat. no. IM304)
  Caution It is imperative to obtain appropriate training from the institutional radiation safety office before receiving radioactivity. Abide by all relevant regulatory rules and use appropriate protection when handling radioactivity. Dispose of the ¹²⁵I-containing radioactive waste according to the institutional radioactive waste disposal guidelines.
• Isoflurane (RxElite Inc., cat. no. NDC60307-120-25)
• U87MG tumor-bearing athymic nude mice (Harlan; see REAGENT SETUP)
  Caution Please obtain appropriate training from the institution regarding animal handling. Animal protocols must be in place before performing animal studies.

Equipment

• HPLC (Dionex)
• Analytical HPLC column (Vydac, cat. no. 218TP54)
• Semi-preparative HPLC column (Vydac, cat. no. 218TP510)
• Freeze-dry system (Labconco Corp.)
• pH paper (EMD Chemicals Inc., cat. no. 9588-3 & 9586-3)
• NAP-10 column (GE Healthcare, cat. no. 17-0854-01)
• Refrigerated microcentrifuge (Eppendorf, 5417R)
• Fluorometer (FluoroMax-3; Jobin Yvon)
• 96-well plate (Thermolab Systems, cat. no. 7417)
• Vacuum manifold (Millipore, cat. no. MAVM0960R)
• Vacuum pump (VWR, cat. no. 89004-272)
• Dry bath incubator (Fisher Scientific, cat. no. 11-718-2)
• Polystyrene culture test tube (Fisher Scientific, cat. no. 14-961-10)
• -counter (Packard)
• GraphPad Prism (GraphPad Software Inc.)
• Glass-bottomed microwell dish (30-mm Petri dish, 14-mm microwell; MatTek, cat. no. P35G-0-14-C)
• Inverted fluorescence microscope (Zeiss)
• Microscope filter set (excitation, 420/40 nm; emission, 705/40 nm; Chroma Technology)
• IVIS imaging system (Xenogen; see EQUIPMENT SETUP)
• Maestro imaging system (CRI Inc.; see EQUIPMENT SETUP)
• Rodent anesthesia system (Summit Anesthesia Solutions)

Reagent setup

• U87MG and MCF-7 cells Culture U87MG cells in DMEM (low glucose) supplemented with 10% (vol/vol) FBS at 37 °C. Culture MCF-7 cells in MEM supplemented with 10% (vol/vol) FBS at 37 °C. The cells should be used when they reach 70–85% confluency. For cell staining, 0.05–0.1 million cells per Petri dish are needed. For cell binding assay, 0.1 million cells per well are needed. Triplicate samples are recommended to obtain statistically meaningful results.
• U87MG tumor model The animals can be used for in vivo imaging studies when the U87MG tumor size reaches about 500 mm³. Typically, we inject 5 10⁶ U87MG cells subcutaneously into athymic nude mice, and it takes 3–4 weeks for the tumor to reach such a size^{31, 32}. The tumor size can be calculated as a b²/2 ('a' represents the longer dimension and 'b' represents the shorter dimension of the tumor).
  Caution Please obtain appropriate training from the institution regarding animal handling and have animal protocols in place before performing animal studies.

Equipment setup

• The Maestro and the IVIS Systems Both systems are quite user friendly once the user has gone through appropriate training of the systems. The key issue is choosing the right excitation/emission filter sets for in vivo imaging. QDs can be excited at any wavelength shorter than their emission wavelength, and the shorter the excitation wavelength, the stronger the absorbance and fluorescence^{1, 2}. However, tissue penetration at shorter wavelength is much poorer. We compared a limited set of excitation filters available in our imaging systems and found that the 590/30 nm excitation gave the best result and the shortest image acquisition time. Excitation at shorter wavelength such as 525/50 nm had dramatically lower fluorescence signal due to poor tissue penetration, although QD705-RGD has much stronger absorbance at this wavelength than at 590 nm. The filter sets we used to acquire the images shown in this protocol are described below. Maestro: excitation, 590/30 nm; emission, 645-nm-long pass. IVIS: excitation, 525/50 nm; emission, 730/80 nm.

Procedure

1. Points from here (point 1) up to and including point 9 are related to Conjugation of QDsMix the following two reagents: QD705 (1 nmol (125 l)) and 4-maleimidobutyric acid N-succinimidyl ester (1 mol in 200 l of borate buffer). Pipette 1–5 l of the reaction mixture and apply it to a pH paper to make sure that the pH is about 8.5.
2. Incubate the mixture in an Eppendorf tube at room temperature (20 °C) for 1 h with gentle shaking.
3. Equilibrate an NAP-10 column with PBS.
4. Load the reaction mixture onto the NAP-10 column; wait until all the mixture is in the column. Add 2 ml of PBS and collect only the deepest-colored fractions (usually 200–400 l). A hand-held UV lamp can help with the visualization.
5. Add 1 mol of RGD-SH (thiolated RGD peptide) dissolved in minimum volume of PBS into the QD705 solution, and adjust the pH to 7.0–7.5 (based on pH paper) with borate buffer if necessary.
   Critical step It is highly recommended to check the purity of the RGD-SH by analytical HPLC before conjugation. If significant amount of disulfide already formed, some TCEP.HCl (1 mg or less) can be added to release the free thiol.
6. Incubate the mixture in an Eppendorf tube at room temperature for 1 h with gentle shaking.
7. Equilibrate another NAP-10 column with PBS.
8. Load the reaction mixture onto the NAP-10 column, wait until all the mixture is in the column, add 2 ml of PBS and collect only the deepest-colored fractions (usually 300–500 l in total).
   Critical step Collect multiple small fractions of the eluent (typically every two drops). Use the most concentrated fraction(s) for future in vivo imaging. Although such procedure may not completely remove all the unreacted RGD-SH, the residual unreacted RGD-SH should not significantly affect future experiments. The most concentrated QD705-RGD fractions should be in the 1–4 M range, suitable for in vivo imaging.
9. Store the collected fractions at 4 °C for later use.Pause Point It is highly recommended to use freshly prepared QD705-RGD conjugate for cell staining or in vivo tumor imaging. We found that QD705-RGD can be stored for up to a few weeks without significant aggregation. However, storing it for several months is not recommended.Troubleshooting
10. Points from here (point 10) up to and including point 20 are related to Characterization of the RGD peptide-conjugated QDsDetermine the concentration of the collected fraction(s) with a fluorometer using serial dilutions of the original QD705 solution as the standard. Typically, a series of concentrations between 0.01 and 10 nM will suffice. Determine the QD705-RGD concentration in triplicate samples using the standard curve. Dilution is necessary, as the collected QD705-RGD fractions are typically between 100 nM to a few M in concentration.
11. Prepare ¹²⁵I-echistatin solution in cell-binding buffer (e.g., 0.5 Ci ml^{- 1}). Typically about 50 l of solution will be needed per well.
    Caution It is imperative to obtain appropriate training previously and abide by all regulatory rules when handling radioactivity.
12. Prepare three stock solutions of different concentration QD705-RGD conjugate (e.g., 2, 20 and 200 nM) in cell-binding buffer. We recommend that parallel experiments using the unconjugated RGD peptide also be carried out (concentrations of stock solutions: 20, 200 and 2 M) for direct comparison of the integrin _v3-binding affinity between QD705-RGD conjugate and the unconjugated RGD peptide. Triplicate samples are recommended.
13. Add 0.02–0.03 Ci of ¹²⁵I-echistatin and appropriate volume of the QD705-RGD stock solution to each well in a 96-well plate. Assuming that the final volume is 200 l per well, the final concentration of the QD705-RGD conjugate should typically range from 10 pM to about 0.1 M.Troubleshooting
14. From a stock of 2 million U87MG cells ml^{- 1} of cell-binding buffer, add 1 10⁵ U87MG cells (50 l) per well, adjust the total volume to 200 l per well with cell-binding buffer and incubate for 2 h at room temperature.
15. Use the vacuum manifold to remove the incubation buffer from the 96-well plate and wash three times with cell-binding buffer (100 l per well).
16. Heat-dry the 96-well plate in the dry bath incubator until the filter membrane is dry. This usually takes about 15 min.
17. Collect the membrane from each well into polystyrene culture test tubes.Pause Point The radioactivity on each membrane can be measured later, as ¹²⁵I has a half-life of 60 d.
18. Measure the radioactivity on each membrane with a -counter.
19. Fit the data by nonlinear regression and calculate the best-fit IC₅₀ (inhibitory concentration of 50%) values for U87MG cells. QD705-RGD should have much lower IC₅₀ values (higher binding affinity) than the unconjugated RGD peptide.
20. Assuming that the QD705-RGD conjugate is functional, they can be used for either cell staining (Option A) or in vivo tumor imaging (Option B).
    1. Cell staining
       1. Pre-seed 0.1 10⁵ U87MG cells (integrin _v3-positive) in each glass-bottomed microwell dish and incubate overnight. Parallel experiments using integrin _v3-negative MCF-7 cells should also be carried out as a control.
       2. Prepare 1 nM solution of QD705 and QD705-RGD in cell-binding buffer (1 ml of each usually suffice).Troubleshooting
       3. Aspirate the cell culture medium from the dish.
       4. Wash the cells 2–3 times with 1 ml of cell-binding buffer (3–5 min each).
       5. Add 150–300 l of the QD705 or QD705-RGD solution to each dish.
       6. Incubate the dishes in the incubator (37 °C) for about 30 min.
       7. Aspirate the staining solution, wash with cell-binding buffer for 3 times (3–5 min each).
       8. Add 500–600 l cell-binding buffer to each dish and examine the cells under a microscope.Troubleshooting
    2. In vivo tumor imaging
       1. Check and make sure that the U87MG tumor size is about 500 mm³.
       2. Set up the Maestro imaging system.
       3. Set up the IVIS imaging system.
       4. Anesthetize the animals using the rodent anesthesia system with isoflurane (2% (vol/vol) isoflurane in 0.2 l min^{- 1} of O₂ flow).
       5. Inject about 200 pmol of QD705-RGD solution in each mouse via tail vein (total volume, 100–200 l). We recommend injecting some PBS (about 50 l) afterward to flush the tail vein.
       6. Scan the animal with either or both imaging systems at serial time points post-injection. We typically choose 10 min, 1 h, 4 h, 12 h and 24 h. Image acquisition time ranges from a few seconds to a few minutes per scan.
       7. For the IVIS imaging system, there is minimal image processing, except adjusting the minimum and maximum value of the color scale. For the Maestro imaging system, spectral unmixing (separating the fluorescence signals of different fluorophores based on the emission spectra^{33, 34}) is needed, which can be done either with the automatic feature embedded in the software or manually. For manual unmixing, the autofluorescence spectrum should be obtained from a normal mouse without QD injection. QD spectrum can be obtained by subtraction of the autofluorescence signal from the mixed signal of a mouse injected with QD705-RGD (e.g., the liver or the U87MG tumor). The unmixed mouse autofluorescence and QD fluorescence images can be color-coded differently and merged.
       8. If needed, harvest the tumor and major organs and image them again using the IVIS and the Maestro system. Ex vivo tissue staining and measuring the tissue homogenate fluorescence can also be carried out to validate the in vivo imaging results.

Timing

Preparation of the peptide-conjugated QDs takes about 3–4 h.
Cell-binding assay usually takes 5–8 h, depending on how many samples are used.
Cell staining typically takes about 2–3 h, excluding pre-seeding the cells the day before and determining the concentration of peptide-conjugated QDs with a fluorometer.

Troubleshooting

Step 9

If the peptide-conjugated QD has been stored for a long period, some aggregation may occur and it will severely interfere with the cell staining and in vivo imaging. For cell staining, the aggregates are very bright, hard to wash away and will overshadow the fluorescence signal resulted from specific RGD-integrin _v3 binding. For in vivo imaging, the aggregates may be trapped in the lung due to the large size. Typically, centrifuge the peptide-conjugated QD at 8,000 r.p.m. at 4 °C in a microcentrifuge for 10 min and only use the supernatant to avoid such problems.

Step 13

Using multiple-channel pipettes can significantly reduce the time of adding the cells and ¹²⁵I-echstatin. However, adding serial concentrations of QD705-RGD or RGD solutions to each well is rather time consuming and prone to error. It takes practice to become proficient. All the ¹²⁵I-containing radioactive waste needs to be collected and disposed under the guidance of the institutional radiation safety office.

Step 20A(ii)

The presence of certain metal ions (e.g., Mn²⁺ and Mg²⁺) in the cell-binding buffer is essential for integrin _v3 staining. The cyclic RGD peptide binds at the major interface between the _v and ₃ subunits and makes extensive contacts with both in a transition metal-dependent mode^{35, 36}. Binding buffer without these ions will result in low counts of the collected membrane. Newly purchased ¹²⁵I-echistatin is recommended for more reliable/accurate results. Using ¹²⁵I-echistatin after more than two half-lives (4 months) will also lead to low counts.

Blocking experiments may also be carried out to confirm the integrin _v3 specificity of the staining. Typically, we use 1 M RGD peptide for blocking. As the blocking solution may cause detachment of the cells from the dish, it is important that the cells are attached well to the glass slip for such blocking studies. Wash the cells with extra care.

Step 20A(viii)

Such fluorescence-based cell staining is semi-quantitative. It is essential that all the cell-staining experiments are carried out using the same experimental procedure and microscope setup. The fluorescence images acquired with the microscope should be displayed at the same pseudocolor scale. Most fluorescence microscope automatically displays the acquired images in 'autoscale' mode and this can be misleading. The image processing software is usually pre-installed on the computer that comes with the microscope. Alternatively, images can also be acquired with a confocal microscope, which may give additional information regarding to what extent the QD705-RGD is internalized upon integrin _v3 binding (integrin _v3 is a membrane-bound receptor and the RGD-binding domain is in the extracellular loop).

Anticipated results

Results from a typical competitive cell-binding assay of RGD and QD705-RGD using ¹²⁵I-echistatin as the radioligand are shown in Figure 3. Compared with the unconjugated RGD peptide, QD705-RGD should have much stronger integrin _v3-binding affinity. Gel electrophoresis has been carried out to confirm the conjugation of proteins on QDs^{27, 28}. However, whether attaching multiple copies of small peptides such as RGD will result in a significantly shifted band has not been tested. QD705-RGD has integrin _v3-specific staining on integrin _v3-positive U87MG cells, whereas QD705 has only minimal nonspecific binding, as shown in Figure 4.

Figure 3: Inhibition of ¹²⁵I-echistatin (integrin _v3-specific) binding to _v3 integrin on U87MG cells by RGD peptide and QD705-RGD (n = 3, mean SD).

The IC₅₀ values for the RGD peptide and QD705-RGD are 255 and 3.3 nM, respectively.

Full size image (22 KB)

Figure 4: Staining of live human breast cancer MCF-7 cells (integrin _v3-negative) and human glioblastoma U87MG cells (integrin _v3-positive) using 1 nM of QD705-RGD.

Staining of U87MG cells with 1 nM of QD705 are also shown. Filter set: excitation, 420/40 nm; emission, 705/40 nm. Magnification: 400. All fluorescence images were acquired under the same conditions and displayed under the same scale.

Full size image (46 KB)

Representative images of U87MG tumor-bearing mice (injected with 200 pmol of QD705 or QD705-RGD) acquired using the Maestro and the IVIS systems are shown in Figure 5. In our study, fluorescence signal resulting from QD705-RGD was clearly visible with both equipments. With the spectral unmixing capability of the Maestro system, it is unambiguous in differentiating the QD705-RGD fluorescence signal from the mouse autofluorescence, which was not possible for the IVIS system used in this study. When using the Maestro system, the typical unmixed emission spectra of QD705-RGD and mouse autofluorescence are shown in Figure 5b.

Figure 5: In vivo imaging of U87MG tumor-bearing mice (yellow arrows) at 6 h post-injection of 200 pmol of QD705-RGD (left) or QD705 (right).

(a) For the Maestro system, the mice autofluorescence is color-coded green and the unmixed QD signal is color-coded red. (b) The 'pure' autofluorescence and QD spectra used for spectral unmixing. (c) Image of the same mouse acquired with the IVIS system immediately after that acquired with the Maestro system shown in a.

Full size image (58 KB)

QD705-RGD mainly targets the tumor vasculature integrin _v3 due to the short circulation half-life²⁰. Ex vivo immunofluorescence staining of the tumor vessels confirmed that the vast majority of injected QD705-RGD does not extravasate from the tumor vessels^{20, 30}. For other vasculature-related targets, the expression level of the target on the vasculature needs to be sufficiently high to observe good tumor contrast. Based on our experience, the QDs we tested are not suitable for targets that are present only on the tumor cells, as the majority of the QD-based conjugate never escaped the circulation to reach the tumor cells.

Author contributions

W.C. and X.C. did the conjugation. W.C. did the imaging experiments and analyzed the data. W.C. and X.C. designed and carried out research. W.C. and X.C. wrote the manuscript.