Synthesis and initial evaluation of radioactive 5-I-α-methyl-tryptophan: a Trp based agent targeting IDO-1

A synthetic approach is established to achieve radioactive 5-I-α-methyl-tryptophan (5-I-AMT).

Abstract

A synthetic approach is established to achieve radioactive 5-I-α-methyl-tryptophan (5-I-AMT). An enzyme based assay demonstrated that 5-I-AMT is a substrate of IDO-1. The in vivo distribution and imaging potential of 5-I-AMT were explored with a B16F10 tumor model using I-124 as the radiotag.

As an immune checkpoint molecule, indoleamine 2,3-dioxygenase (IDO-1) is one of the first and rate limiting enzymes in the kynurenine pathway that metabolizes tryptophan (Trp).1 In fact, elevated IDO-1 expression not only causes Trp depletion (leading to immune cell deactivation), but also results in the production of kynurenine, which selectively impairs T cell growth and survival. Due to its immunomodulatory role in cancer progression, IDO-1 has become an interesting target from both a therapy and an imaging point of view.2 The clinical importance of the IDO-1 pathway is reflected upon the clinical development of specific IDO-1 inhibitors which, in combination with PD-1 inhibitors, have shown significantly higher antitumor activity3 across various solid tumors in early clinical studies compared with either agent administered alone.4 This fast advancement prompts the urgent need to answer the following questions: which patients should be treated with the PD-1 + IDO-1 combination therapy, when should the IDO-1 inhibitor be given during PD-1 treatment, and what IDO-1 inhibitor dosage is sufficient to reduce IDO-1 activity? Since repetitive biopsies of the same patient are impractical, a non-invasive method that monitors IDO-1 activity in vivo is highly preferred.

Previously, the PET tracer α-¹¹C-methyl-l-tryptophan (¹¹C-AMT, 1) has been developed to study tryptophan metabolism in vivo including brain serotonin synthesis, epileptic foci, epileptogenic tumors, and primary and metastatic brain tumors.5 Major limitations of ¹¹C-AMT are the sensitive reaction conditions involved with its synthesis and the short half-life of ¹¹C (20.36 minutes). These drawbacks prevent its commercial distribution and widespread application in the clinic. Moreover, ¹¹C-AMT could be metabolized through both the serotonin and kynurenine pathways, which makes it a less specific agent for IDO-1. In order to address the aforementioned issues, ¹⁸F labelled Trp derivatives have been developed to monitor IDO-1 activity in vivo. Recently, we demonstrated that the selectivity between the IDO-1 pathway and the Tph pathway could be improved by replacing the hydrogen atom at position 5 of AMT with fluorine (Scheme 1).6 Because iodine radioisotopes are readily available for biomedical applications, in this report, we explore the synthesis of radio-iodinated AMT agents that could potentially be used for both cancer imaging and therapy targeting the IDO-1 pathway.

Scheme 1. Trp based agents targeting the IDO-1 pathway.

There are four commonly used iodine radioisotopes in the biomedical field: ¹²³I (for SPECT imaging, t_1/2 = 13.22 hours), ¹²⁴I (for PET imaging, t_1/2 = 4.18 days), ¹²⁵I (for therapy and SPECT imaging, t_1/2 = 59.4 days), and ¹³¹I (for therapy and SPECT imaging, t_1/2 = 8.02 days). Here, we aim to radiolabel the 5-position of AMT with ¹²⁴I, whose relatively long half-life would allow for imaging late time points. The obtained biodistribution and excretion profile of 5-I-AMT would allow us to evaluate whether it could be used as a theranostic (therapy and diagnostic) agent when labelled with other radio-iodines (Scheme 1).

Our initial attempt to synthesize ¹²⁴I-5-I-AMT involved incorporation of radioiodine via the iododestannylation of stannane 4 using hydrogen peroxide and acetic acid (Scheme 2a).7 Incorporation of radioiodine into the substrate was observed by radio HPLC analysis of the crude reaction mixture. However, the subsequent deprotection conditions decomposed the product. Efforts to synthesize the free, unprotected amino acid with the trimethylstannane group at position 5 failed. We then switched to the copper mediated isotope exchange approach as shown in Scheme 2b.8 Radioiodine is incorporated into the unprotected amino acid without the need to perform protection/deprotection. This route provided [¹²⁴I]-5-I-AMT in 95% isolated yield, with a specific activity of approximately 0.74 mCi μmol^–1.

Scheme 2. Reaction routes to synthesize 5-I-AMT.

With this simple and efficient method to generate radioiodinated AMT, we then performed an enzyme based assay to test whether 5-I-AMT is a substrate for the IDO-1 enzyme, and compared the result with that of the natural substrate tryptophan. As shown in Fig. 1A, the absorbance at 490 nm for the reaction containing l-Trp or I-AMT was 0.241 ± 0.009, 0.161 ± 0.003 (n = 3), while the control reaction system (no substrate) had very low absorbance (0.076 ± 0.002). These results suggested that I-AMT is a substrate for IDO-1, and the radiolabeled I-AMT could therefore be used for IDO-1 imaging.

Fig. 1. (A) IDO-1 enzyme assay demonstrated that I-AMT is a substrate of IDO-1. (B) PET of ¹²⁴I-AMT in B16F10 xenografts in C57BL6 mice.

With the promising in vitro data on hand, [¹²⁴I]5-I-AMT was then evaluated in vivo using PET in B16F10 tumor-bearing C57BL6 mice. Similar to AMT, ¹²⁴I-AMT is a small molecule that should be cleared from blood relatively fast. Indeed, we did not observe obvious blood activity at 0.5 h post injection. The elevated enzyme activity in the tumor should increase its tumor retention: after intravenous injection of ¹²⁴I-AMT, B16F10 tumor xenografts can be clearly visualized with good tumor-to-background contrast. The representative PET images are shown in Fig. 1B. For the B16F10 tumor, the uptake was 3.24 ± 1.20, 3.42 ± 0.97, and 0.83 ± 1.03% ID per g at 0.5 h, 1.4 h, and 20 h after injection, respectively. For an unknown reason, the variation at the 20 h time point is much larger than expected. Some animals demonstrated high uptake at 20 h p.i., while others showed low tumor retention. Compared with AMT analogs that are labeled with ¹¹C or ¹⁸F, the uptake is roughly around the same range at early time points. However, a side-by-side comparison on the same tumor model is needed before a conclusion is drawn. The kidney also showed a high uptake of ¹²⁴I-AMT, which was 27.93 ± 5.89, 22.52 ± 4.83, and 3.15 ± 3.32% ID per g at 0.5 h, 1.4 h, and 20 h after injection, respectively. The uptake of ¹²⁴I-AMT in liver was comparable to that in the B16F10 tumor, while muscles only showed minimal tracer accumulation. The ex vivo tissue biodistribution results were consistent with PET quantification. No apparent thyroid uptake was seen at all time-points studied, indicating that de-iodination did not play a major role in vivo. In fact, an in vitro stability study also demonstrated that ¹²⁴I-AMT is stable up to 20 h post incubation. In cultures of PBMCs, IDO-1 was found predominantly in monocytes by immunohistochemistry. Reverse transcriptase polymerase chain reaction analysis showed that IDO-1 mRNA was expressed in T lymphocytes, B lymphocytes and natural killer (NK) cells, and that expression was increased upon activation with interferon-γ.9 To minimize the impact of IDO-1 expressed on immune cells, and also to further test the specificity of the IDO-1 targeting efficacy of ¹²⁴I-AMT, we applied the IDO-1 positive B16F10 xenograft model in NSG mice, in which mature B cells, mature T cells, and natural killer cells are absent and dendritic cells and macrophages are defective.10 After the tumor reached 0.6 cm in diameter, serial static small animal PET scans were performed in the B16F10 tumor bearing NSG mice and B16F10 xenograft bearing C57BL6 mice. After intravenous injection of ¹²⁴I-5-I-AMT, the B16F10 tumor can be delineated clearly in both mice strains. The quantitative ex vivo tissue biodistribution results are shown in Fig. 2B. The accumulation of ¹²⁴I-AMT was comparable in major organs, including the heart, lung, liver, pancreas, spleen, stomach, small intestine, kidneys, muscles and bone (n = 3, p > 0.05). However, ¹²⁴I-AMT accumulation was significantly higher in B16F10 developed in C57BL6 mice (3.66 ± 0.69% ID per g) than that in NSG mice (1.55 ± 0.86% ID per g) (p < 0.05, n = 3) based on the biodistribution study. Representative PET images are shown in Fig. 2A.

Fig. 2. Comparison of I-AMT distribution using (A) PET and (B) biodistribution in B16F10 established in C57BL6 or NSG mice.

To confirm that the reduced ¹²⁴I-AMT uptake indeed reflects IDO-1 expression in tumor tissues, we compared IDO-1 expression in B16F10 tumors grown in C57BL6 or NSG mice. As shown in Fig. 3, the B16F10 tumor grown in NSG mice had a less IDO-1 positive signal as compared with that grown in C57BL6 mice, which can not only have IDO-1 expressed on tumor cells, but also have IDO-1 expressed on immune cells. We did not observe different IDO-1 expression between spleens of C57BL6 or NSG mice. We would like to point out that the specific activity of ¹²⁴I-AMT is relatively low compared with no-carrier added PET agents. However, it may not affect the overall distribution of the agent considering it is a substrate based probe. Nonetheless, if the high specific activity is achieved in the future, the impact caused by this factor should be investigated further.

Fig. 3. Immunohistochemical staining of IDO-1 expression (the scale bar in the tumor section represents 100 μm and the scale bar in the spleen section represents 200 μm).

Conclusion

To study tryptophan metabolism in vivo, several radionuclides have been used to produce radiolabeled tryptophan derivatives. In this study, we established a simple method to produce radioactive [¹²⁴I]5-I-AMT. Matching of the radionuclide half-life to the pharmacokinetic profile of tryptophan analogs is an important criterion for radionuclide selection. In addition, practical radiochemical labeling of tryptophan without compromising its reactivity with IDO-1 was also a key criterion. Taking these considerations, isotope ¹²⁴I was chosen in this study, which allowed us to not only use PET to study the distribution of 5-I-AMT, but also explore the uptake at late time points due to its 4.18 day physical half-life. Indeed, [¹²⁴I]-5-I-AMT PET imaging demonstrates that the agent has prominent uptake within the first day, but cleared at the 24 h time point. The data obtained here suggest that the use of [¹³¹I]5-I-AMT as a therapeutic agent may be limited to a certain extent by the mismatched half-life of ¹³¹I (t_1/2 = 8.02 days) and relatively fast clearance of 5-I-AMT. In contrast, [¹²³I]-5-I-AMT may deserve further exploration as a SPECT agent in future studies since the half-life of ¹²³I (13.22 h) pairs well with the clearance profile of 5-I-AMT.

This work was supported by NIH (1R01CA233904-01), P30-CA016086-35-37 from the National Cancer Institute, and UNC Radiology Department and BRIC. Small animal PET results were obtained with instrumentation supported by NIH grants 1S10OD023611.

Conflicts of interest

There are no conflicts to declare.