Combinatorial peptide libraries, which were first introduced in early 1990s,1-4 have become an important tool for drug discovery.5-8 The basic principle in the application of combinatorial peptide libraries in drug discovery is that lead compounds can be identified by simultaneous, systematic screening a large number of peptides. The same principle also can be applied to small organic molecules.9-11
A fundamental limitation to current combinatorial technique is lack of sensitive, quantitative and dynamic screening in vivo noninvasively. The seemingly perfect lead compound, from a large numbers of candidate compounds in vitro, may not necessarily be specific in targeting the intended tissue due to its interference with unknown biomolecules and metabolism in tissues.
Radiolabelling method, as a sensitive and simple approach for assessment of drug distribution and bioelimination, has become a choice for targeting tissues in drug research.12 Theoretically, libraries of radiolabelled compounds make high throughput screening possible in vivo. Since the diagnostic and therapeutic efficiency of a drug depends largely on its distribution in target tissues both selectively and accumulatively, screening of libraries in vivo is believed to be a practical tool for basic research and drug discovery. In this article, the iterative and pooling strategy13 was demonstrated successfully in our screening study with a library of 99mTc tripeptoid complexes. Using sublibraries of 99mTc tripeptide complexes provided a novel model for both random screening and directed screening in tissue targeting, lead generation and optimization.
Material, cell line and animal
Twenty 9-fluorenylmethoxycarbonyl (Fmoc) natural L-aminoacids (purity>98.5%) and O-benzo-triazol-1-yl-N, N, N', N'- tetrametyluranium tetrafluoroborate (TBTU, purity>99%) were obtained from GL Biochem, Shanghai, China and Rink Amide-MBHA resin was obtained from Tianjin Nankai Hecheng Science & Technology Co., China. Other reagents, solvents and culture media were purchased from Sigma-Aldrich and Fluka, USA. Female SCID/beige mice and female Sprague-Dawley (SD) rats of 6 to 8 weeks old were purchased from Fudan University, Shanghai, China. Cell line A549 of human lung cancer was purchased from Shanghai Institute of Cells and cultured in RPMI 1640 medium containing 10% foetal bovine serum under a fully humidified atmosphere containing 5% CO2 at 37°C. The same number of cells (2 ×105) was subcutaneously transplanted into the chest wall of each mouse.
The iterative and pooling strategy of C-amide tripeptide library
The C-terminal amide library was synthesized on Rink Amide-MBHA resin14 in a OXX → O1OX → O1O2O → O1O2O3 iterative and pooling protocol, where O represents one of the 20 natural L-aminoacids, O1—O3 each represents a specified L-aminoacid in a particular position, X represents any of the L-aminoacids (with the exception of Cys). If the screening signals generated from labelled O1O2O were weak, the C-terminal L-aminoacid made less contribution to individual active sequences and the iterative route was changed. After labelled with 99mTc, 8,000 99mTc complexes of different structures were in the library. The proposed structure of the 99mTc tripeptide complexes and iterative and pooling protocol in vivo are shown in Fig.1.
Synthesis of C-amide library and peptides
The mixture of 19 L-aminoacids was prepared as described by Houghten RA.2 The library and tripeptides were synthesized on Rink Amide-MBHA resin in MiniKan reactors (IRORI products, San Diego, USA) according to a standard method of Fmoc solid phase peptide synthesis.2,14 The coupling efficiency was monitored by staining resin with 0.005% bromophenol blue in DMF. The quality of libraries and peptides were analysed by Dansyl-Edman sequencing. The peptide products were dried under vacuum without further purification, stored at —25°C and used within 3 days.
Preparation of 99mTc-libraries and 99mTc-peptoids
The C-amide tripeptides are a class of N4triligands.15 Similar to the complexes of technetium with tetrapeptides,16 an oxotechnetium (V) group was proposed to bind the N4-triligand via two deprotonated peptide nitrogen atoms, one deprotonated amide and the free electron pair of one amine nitrogen atom to form a structure of 99mTc tripeptide complex (Fig. 1). The 99mTc labelling procedure was modified from Vanbilloen HP.16 Briefly, 100 to 300 μg of the C-amide tripeptide sublibrary (or single tripeptide) was dissolved in 0.5 ml 0.1 mol/L NaOH (pH=13), 0.4 ml of this solution was then injected into a 10 ml sterile vacuum vial. A newly prepared solution of SnCl2 · 2H2O (100 μg in 15 μl 0.05 mol/l HCl) was added, immediately followed by the addition of 0.1 ml of eluate (generator from Syncor International Corporation, Shanghai, China) containing 37 to 370 MBq (1 to 10 mCi) 99mTc in the form of sodium pertechnetate (NaTcO4). The mixture was heated for 10 minutes in a boiling water bath then cooled to room temperature. For purification of the labelled products, a 20 channel thin layer chromatography separation was performed on two GF254 silica gel plates backed with terylene (20 cm × 10 cm each) to deal simultaneously with 20 samples. The impurity of [99mTc]-TcO4 - (Rf = 0.95 to 1.0) was first removed using propanone as the mobile phase. A mixture (cyanomethane: propanone: normal saline = 1.5:1:1) was used as the second mobile phase to elute the labelled products and leave the impurity of colloidal 99mTc at the application point (Rf = 0.0 to 0.1). The labelling products were recovered from saline after removal of silica gel by centrifuging at 2000× g for 5 minutes. The solution of labelled products in saline was passed through a 0.22-μm filter membrane before in vivo administration.
Labeling efficiency of 99mTc-libraries
To analyse the labelling discrepancy of individual peptides in the library, sublibraries of OXX, XOX, and XXO were synthesized and were labelled with 99mTc. The percentage of [99mTc]-TcO4- and colloidal 99mTc was measured by paper chromatography.16
Biodistribution in animal
After injection of labelled sublibraries or a single complex, 74 to 185 kBq per animal via tail vein, all animals were kept under normal conditions. SD rats were sacrificed by administration of excessive thiopental sodium 1.5 hours after injection. Tissues were removed, weighed, washed with cold saline, and counted for 99mTc activity in a γ -counter. The tissue counts were corrected for background radiation excluding 99mTc physical decay as the injected dose was also counted concurrently. Results were expressed as percent of injected dose per gram tissue (%ID/g). For noninvasive studies, rats and tumour bearing mice received i.v. high dose of99mTc sublibraries or single 99mTc tripeptide complex at 3.7 MBq per animal. Imaging studies were performed with a γ-camera at a preset time after injection. To keep animals under anaesthesia for 1 to 2 hours during imaging, thiopental sodium (70 mg/kg body weight) was administered intraperitoneally just before imaging studies.
RCSin vivo:iterative/pooling and lead identification
RCS in vivo was carried out by iterative/pooling as above description (Fig. 1). The typical procedure, used sublibraries in order 99mTc OXX→99mTc O1OX→99mTc O1O2O→99mTc O1O2O3. Twenty animals each received i.v. 1 sublibrary of 99mTc OXX (99mTc AXX,99mTc DXX, 99mTc EXX 99mTc YXX, each sublibrary including 400 99mTc tripeptide variations of the same O1), and tissue biodistributions of those libraries were recorded and analysed. Of the 99mTc OXX sublibraries, 99mTc RXX had highest %ID/g value in kidneys (Fig. 5A), thus it went on the next round of screening of 20 sublibraries of 99mTc O1OX (99mTc RAX, 99mTc RCX, 99mTc RDX, 99mTc REX99mTc RYX for renal targeting, Fig. 5B, again using 20 animals each receiving 1 99mTc O1OX sublibrary of 20 variations). Accordingly, the most effective 99mTc O1O2O3 (99mTc RES for renal targeting, Fig. 5C), was identified after the final round screening of 20 99mTc O1O2O (99mTc REA, 99mTc REC, 99mTc RED, 99mTc REE 99mTc REY, each animal receiving 1 99mTc tripeptide). Considering the statistical error, tissue biodistribution studies were repeated twice for each best candidate of labelled sublibraries or labelled peptides.
Planar images of the animals were acquired with a dual head γ -camera (Multi-Spect-2, SIEMENS) equipped with a low energy, high resolution and parallel hole collimator. Energy peak was set at 140 keV±10% for 99mTc. Animals were placed in direct contact with collimator covered with plastic wrap. Cumulative counts were at least 100 000 for each animal. Images were recorded on a 256×256 matrix with a 3.2 zoom.
Labeling discrepancy and potential screening bias
The labelling discrepancies of 99mTc libraries are shown in Fig. 2. The mean labelling yields of 99mTc OXX, 99mTc XOX and 99mTc XXO sublibraries were (86.59 ± 5.31)%, (78.66 ± 7.04)%, and (64.13 ± 9.64)% respectively. The major uneven labelling was found in C-amide-terminal of 99mTc XXO, indicating that the 99mTc peptides, which containing tryptophan, proline, histidine and phenylalanine at C-amideterminal (Fig. 2), might be underestimated during screening in mixed condition. In mixed condition, another potential screening bias (underestimation) comes from construction of tryptophan or proline at position 2 of 99mTc peptides (Fig. 2). There is no major bias problem at N-terminal (Fig. 2). To minimize the bias caused by uneven labelling of tripeptides in mixed condition, the iterative screening should preferentially start from C-amide-terminal. Unfortunately, the radio signal in tissues was weak if screening in vivo started from 99mTc XXO. We decided therefore to adopt the iterative screening from 99mTc OXX to 99mTc 1O2O3 (see the methods).
Directed screening: optimization of renal function agent
Directed screening (or chemical analoging), where the objective is to evaluate closely related structural analogs of a lead molecule, establish SAR (structure activity relations), and optimize biological potency and other pharmaceutically relevant properties.17 In the case of 99mTc tripeptide complexes, the labelled structures have the property of choice for renal function studies, since 80.79% activity of 99mTc XXX library appeared in urine 1.5 hours after intravenous injection. We wondered if a 99mTc tripeptoid, with better renal excretion characteristic, could be optimized from all of the 8000 99mTc tripeptoid candidates. With the assistance of in vivo RCS, we accomplished this work using only 66 rats within 15 work days. The results are shown in Fig. 3. After screening of 20 sublibraries of 99mTc OXX, 99mTc DXX was found the best candidate (Fig. 3A). Another 20 sublibraries of 99mTc DOX were then prepared for the next screening, and 99mTc DSX was demonstrated the best candidate (Fig. 3B). Finally, the urinary excretion of 99mTc DSG was found the best among 20 single tripeptide of 99mTc DSO (Fig. 3C). In a renal imaging study, 99mTc DSG can be quickly excreted through rat kidneys within 12 minutes (Fig. 3D).
Random screening: identification of tissue targeting leads
Random screening is where the task is to identify a novel lead compound in the absence of any structural or mechanistic information about the biological target macromolecule17 or in the absence of any information about the biological targets in a certain tissue as in vivo screening. Though the majority of labelled peptides were excreted through kidney without specific tissue distribution, a few tissue specific, 99mTc tripeptoids had been identified from the random screening using the same procedure described above. They are 99mTc RPA,99mTc VIG and 99mTc RES, which possessed selective distribution in kidney, liver and tumour respectively (Table 1 and Fig. 7).
Initial and final random screening
Biodistribution in SD rats of 99mTc OXX libraries is shown in Fig. 4. Each 99mTc OXX sublibrary (99mTc AXX, 99mTc DXX, 99mTc EXX 99mTc YXX) contains 400 variations of 99mTc complex (Fig. 4A, the initial screening). The radioactivity of labelled sublibraries in target and nontarget tissues was detected and analysed simultaneously. After initial random screening, second round of screening started from 99mTc RXX (for kidney) and 99mTc IXX(for liver). Biodistribution of 99mTc REO is shown in Fig. 4B, which presents the final screening of lead for kidney imaging. Comparing with the initial screening, the best lead of 99mTc RES had by far the highest distribution in kidney with low background activity (Fig. 4B).
Identification of kidney targeting leads
Random screening of 99mTc tripeptoids targeting kidney is shown in Fig. 5. Following the iterative/pooling procedure of 99mTc OXX→ 99mTc ROX→ 99mTc RE(P)O, 99mTc RES and 99mTc RPA were finally identified. Fig. 5 shows that the accumulation of 99mTc RES in kidney is highly structure dependent. The %ID/g of 99mTc RES in kidney was (16.92 ± 3.13)%, 282-fold, 20-fold, 15-fold, and 12-fold higher than that of 99mTc REC,99mTc RET, 99mTc RSE, and 99mTc RDS respectively.99mTc RPA was unstable in vivo, the free 99mTc activity from catabolism of 99mTc RPA accounted for about 27.5% of total activity in serum 1.5 hours after injection. The rat images of 99mTc RES and 99mTc RPA are shown in Fig. 5D.
Identification of liver targeting lead
After the first round screening of 99mTc OXX, high liver activities were detected among the sublbraries of 99mTc IXX, 99mTc PXX, 99mTc QXX and 99mTc RXX (Fig. 6A). Only two sublibraries of 99mTc IGX and99mTc IIX showed strong signals in the liver during second round screening (Fig. 6B). Weak signals appeared however in the third round screening of99mTc IGO. Additional round screening of 99mTc OIG was therefore carried out and 99mTc VIG found to be the most active (Fig. 6C). A relatively sustained 99mTc VIG activity for liver with very low tissue background is shown in Fig. 6D. The %ID/g of 99mTc VIG in liver was highly structure dependent (Table 2).
Identification of tumour targeting lead
In order to save experimental animals and monitor the bioelimination of labelled libraries as well, noninvasive screening was applied to identify 99mTc tripeptoids avidity to tumours. The imaging study of tumour bearing mice was performed 4 hours after injection of labelled libraries (or individual peptides). After in vivo screening of 99mTc EXX, 99mTc REX and99mTc REO, it turned out that 99mTc RES was the best for tumour targeting (Fig. 7).
Optimisation of lead Design
The identification of 99mTc VIG indicated that side chain methylation at positions 1 and 2 was crucial for liver accumulation of 99mTc tripeptoids. To demonstrate reliability of the information, liver biodistribution in rats of the various 99mTc tripeptoids, with different side chain methylation at defined positions, was compared (Table 2). 99mTc GGG without side chain methylation shows the lowest liver radio uptake. The presence of methyl substituent on the first, second or third aminoacid separately has little influence on accumulation in the liver. The presence of methyl substituent on the both first and second aminoacid however favoured considerably accumulation in the liver of 99mTc tripeptoids.
Although new lead discovery in pharmaceutical industry is more effective having benefited from large combinatorial libraries of peptide and nonpeptide molecules produced using high throughput screening paradigms,8,18,19 much less success of in vivo tissue targeting leads could be expected due to the interference of background binding and metabolism of leads in tissues. We therefore designed a practical approach, which involved labelling libraries with radioisotope and screening in vivo, to identify rapidly leads for target tissues.
By means of γ emission counting, radiolabelling technology can provide a noninvasive approach to trace simultaneously a large number of compounds in vivo, sensitively and quantitatively. After injection of the radiolabelled libraries into experimental animals, the radio signals of tissue distribution and bioelimination of the libraries were recorded by γ counting or imaging. For in vitro affinity screening, washing is required to remove unrelated compounds or free biomolecules so that the active compound (s) can be identified. For in vivo screening, compounds without affinity to tissues either are in blood pool or are excreted into urine and/or intestine after a period of in vivo metabolism, just as in the washing procedure used in vitro. The signals of labelled compounds in target tissues could therefore be recorded directly.
With a reasonable library design and appropriate building block selection, a variety of radioisotopes (for example, 13C, 18F, 125I, 111In and 99mTc) can be used to label peptide or small nonpeptide libraries. In this study using 99mTc labelled sublibraries, screening in vivo was demonstrated successfully in both optimization of leads and identification of compounds for tissue targeting.
In the case of directed screening, structural libraries of lead analogue could be synthesized if the lead structure was known and had highly selective affinity to a defined target protein or tissue. After directed screening in vivo, we found 99mTc DSG was the most promising renal tracer among the 8000 99mTc complexes, although most of these were excreted into urine too. Moreover, we had found the tissue accumulation of 99mTc tripeptoid leads was highly structure dependent. With assistance from the structure information provided by in vivo screening, the design of tissue targeting drug could thus become feasible and more interesting.
Another advantage of in vivo screening is the direct identification of compounds against target tissues without background knowledge of what the particular molecular targets might be, ie, the random screening. Any biomolecule in the tissue of interest could turn out to be the target of the compound(s) in a defined labelled library. For instance, 99mTc VIG and 99mTc RES were demonstrated to have a specific tissue target to liver and tumour respectively, though their exact molecular mechanisms are unknown. This type of screening is therefore totally randomized. Using labelled compounds as the probes, understanding of relevant mechanisms could become more effortless.20,21 For instance, tumour angiogenesis may contribute to the accumulation of99mTc RES, as the 99mTc RES activity in tumour could be in part replaced by excessive "cold" RGD (data are not presented). During the process in which a99mTc labelled lead was identified, the nontarget tissue binding and bioelimination of labelled libraries were monitored successfully so that the background activity was limited to the lowest level. Considering that the molecular targets for radiotracer studies are so limited currently,22,23 it became more important and practical for the simple, low cost, quick, and noninvasive tool of in vivo screening to be applied in the discovery of useful radiotracers.
The inherent limitations of in vivo screening are worth mentioning. As a result of possible interference from weakly active compounds and uneven labelling among individual compounds, the most active sublibrary for tissue targeting may not contain the single most active individual compound. In this case, as we identified with 99mTc VIG for liver targeting, additional iterative screening should be carried out so that the most valuable compound could be picked out from a heap of "straw". Of course, the uneven labelling problem could be solved if the sublibraries are conjugated with a bifunctional ligand, HYNIC or MAG3, for instance. In addition, when applied to screening of large libraries, the radio signals can become too weak to trace the huge number of individually labelled compounds.
1. Scott JK, Smith GP. Searching for peptide ligands with an epitope library. Science 1990;249:386-390.
2. Lam KS, Salmon SE, Hersh EM, Hruby VJ, Kazmierski WM, Knapp RJ. A new type of synthetic peptide library for identifying ligand-binding activity. Nature 1991;354:82-84.
3. Houghten RA, Pinilla C, Blondelle SE, Appel JR, Dooley CT, Cuervo JH. Generation and use of synthetic peptide combinatorial libraries for basic research and drug discovery. Nature 1991;354:84-86.
4. Devlin JJ, Panganiban LC, Devlin PE. Random peptide libraries: a source of specific protein binding molecules. Science 1990;249:404-406.
5. Irving MB, Pan O, Scott JK. Random-peptide libraries and antigen-fragment libraries for epitope mapping and the development of vaccines and diagnostics. Curr Opin Chem Biol 2001; 5:314-324.
6. Golebiowski A, Klopfenstein SR, Portlock, DE. Lead compounds discovered from libraries: part 2. Curr Opin Chem Biol 2003;7:308-325.
7. Sehgal A. Recent developments in peptide-based cancer therapeutics. Curr Opin Drug Discov Devel 2002;5:245-250.
8. Gordon EM, Barrett RW, Dower WJ, Fodor SP, Gallop MA. Applications of combinatorial technologies to drug discovery. 1. Background and peptide combinatorial libraries. J Med Chem 1994; 37:1233-1251.
9. Bunin BA, Plunkett MJ, Ellman JA. The combinatorial synthesis and chemical and biological evaluation of a 1,4-benzodiazepine library. Proc Natl Acad Sci U S A 1994;91:4708-4712.
10. Wood WJ, Huang L, Ellman JA. Synthesis of a diverse library of mechanism-based cysteine protease inhibitors. J Comb Chem 2003;5:869-880.
11. Thompson LA, Ellman JA. Synthesis and applications of small molecule libraries. Chem Rev 1996;96:555-600.
12. Port RE, Wolf W. Noninvasive methods to study drug distribution. Invest New Drugs 2003;21:157-168.
13. Pinilla C, Appel JR, Blondelle SE, Dooley CT, Eichler J, Ostresh M, et al. Versatility of positional scaning synthetic combinatorial libraries for the identification of individual compounds. Drug Dev Res 1994;33:133-145.
14. Fields GB, Noble RL. Solid phase peptide synthesis utilizing 9-fluorenylmethoxycarbonyl aminoacids. In J Peptide Protein Res 1990;35:161-214.
15. Boussard C, Doyle VE, Mahmood N, Klimkait T, Pritchard M, Gilbert IH. Design, synthesis and evaluation of peptide libraries as potential anti-HIV compounds, via inhibition of gp120/cell membrane interactions, using the gp120/cd4/fab17 crystal structure. Eur J Med Chem 2002;37:883-890.
16. Vanbilloen HP, Bormans GM, DE Roo MJ, Verbruggen AM. Complexes of technetium-99m with tetrapeptides, a new class of 99m
Tc labelled agents. Nucl Med Biol 1995; 22:325-338.
17. Gordon EM. Strategies in the design and synthesis of chemical libraries. In: Gordon EM, Kerwin JF Jr, editors. Combinatorial Chemistry and Molecular diversity in Drug Discovery. New York: John Wiley and Sons; 1998:17-38.
18. Moos WH, Green GD, Pavia MR. Recent advance in the generation of molecular diversity. Annu Rep Med Chem 1993;28:315-324.
19. Otto S, Furlan RL, Sanders JK. Dynamic combinatorial chemistry. Drug Discov Today 2002;7:117-125.
20. Chatziioannou AF. Molecular imaging of small animals with dedicated PET tomographs. Eur J Nucl Med Mol Imaging 2002;29:98-114.
21. Wust F. Use of [99m
Tc]technetium-labeled steroids as probes for steroid hormone receptors. Methods Mol Biol 2001;176:133-143.
22. Virgolini I, Traub T, Novotny C, et al. New trends in peptide receptor radioligands. Q J Nucl Med 2001; 45:153-159.
23. Kwekkeboom D, Krenning EP, de Jong M. Peptide receptor imaging and therapy. J Nucl Med 2000; 41:1704-1713.
Keywords:© 2006 Chinese Medical Association
99mTc labeled; tripeptide libraries; in vivo; random screening; directed screening