2.1 Imaging of TAMs

The macrophage mannose receptor (MMR, also termed CD206), is a member of the C-type lectin family, and represents a promising target to visualize TAMs. Many MMR-targeting probes have been developed using dextran nanoparticles (DNPs), including ⁶⁴Cu-labeled/mannosylated liposomes [31], ¹⁸F-macroflor [65], ⁸⁹Zr-DNPs [66], and ⁶⁴Cu-labeled carbohydrate-derived DNP macrin (Figure 2a) [32]. Because drugs can be encapsulated into DNPs [67, 68], these nanoparticles are useful to deliver drugs to tumor lesions, thereby achieving theranostic effects. However, mannose and its analogs are not specific for CD206 and have cross-reactivity with other mannose receptors such as CD209 [69], prompting the recent development of the CD206-targeting nanobody (Nb). Two groups isolated and characterized two anti-CD206 nanobody-based tracers (^99mTc-labeled α-MMR Nb and ¹⁸F-fluorobenzoate [FB]-anti-MMR 3.49) (Figure 2b) [33, 34]. Both tracers show high retention in tumors and minimal uptake in normal organs/tissues, which is consistent with their specific recognition of TAMs in vivo. Their rapid elimination from circulation enables imaging as early as a few hours after injection, which is in direct contrast to mAbs that require days between injection and imaging [33, 34]. Considering that Nbs outperform mAbs in terms of solubility, stability, and affinity, they have a greater potential for CD206-targeted TAM imaging [70].

CD163 (hemoglobin scavenger receptor) is restricted to the monocyte–macrophage lineage and upregulated in TAMs. Eichendorff et al. developed an anti-CD163 antibody ([⁶⁸Ga]ED2) and performed PET/CT imaging in rats with collagen-induced arthritis [36]. Surprisingly, no studies have attempted tumor imaging by targeting CD163.

Folate receptor beta (FR-β) is upregulated in TAMs. Its tracers include aluminum ¹⁸F-labeled 1,4,7-triazacyclononane-N,N′,N″-triacetic acid conjugated folate (¹⁸F-FOL) [37], and ¹⁸F-AzaFol [38]. These tracers show specific uptake in macrophages expressing FR-β. However, it should be noted that FR-β is also expressed on certain human cancer cells, which might hinder the clinical applicability of these tracers for macrophage-specific imaging [71]. Additionally, radiolabeled nanoparticles and liposomes can be used to visualize TAMs by their phagocytotic capacity, such as ⁸⁹Zr-conjugated high-density lipoprotein (⁸⁹Zr-PL-HDL and ⁸⁹Zr-AI-HDL) (Figure 2c) [35]. Although F4/80 and CD11b are well-known macrophage markers, they are unsuitable imaging markers because the human analog of the former is an eosinophil-specific marker and of the latter is a pan-monocytic marker expressed on TAMs, MDSCs, and TANs [40, 69].

The tumor-permissive and immunosuppressive characteristics of TAMs have fueled interest in therapeutically targeting these cells. Tracers enable noninvasive visualization of TAMs by nuclear imaging and can deliver therapeutic radionuclides to the TME. However, further efforts should be made to identify specific markers and develop specific probes.

2.2 Imaging of TANs

Although TANs play a major role in the TME, tracers to visualize TANs are lacking. Some examples include the NE 680 FAST imaging agent, which emits fluorescence after cleavage by NE, a biomarker of TANs [41]. An amphiphilic semiconducting polymer conjugated with a hemicyanine (hemi-Cy) dye (SPCy) has been synthesized to detect NE. It passively targets the tumor and uncages the hemi-Cy in response to NE, leading to enhanced near-infrared fluorescence and photoacoustic signals of hemi-Cy [42]. Cinnamoyl-F-(D)L-F-(D)L-F (cFLFLF), which targets the formyl peptide receptor on inflammatory cells, including neutrophils, remains to be evaluated for TAN imaging [43]. However, because unresolved inflammation is common during tumor progression, cFLFLF might be a promising probe to evaluate local inflammation in the TME. Collectively, imaging of TANs is imperative for cancer immunotherapy but remains challenging. However, significant attention has been garnered to unravel its multifunctional capability in heterogeneous subpopulations, during which novel markers such as CCL4 have been characterized with the potential to develop new tracers [72].

2.3 Imaging of MDSCs

Because MDSCs share the same marker, CD11b, a ^99mTc-labeled anti-CD11b mAb has been developed to image colon cancer, and tumor nodules as small as 3 mm could be clearly identified [44, 45]. Hoffmann and colleagues tracked the dynamic changes of MDSCs in vivo using a ⁶⁴Cu-modified CD11b-specific mAb [45]. Additionally, a [⁸⁹Zr]anti-CD11b Ab with high specificity was also developed recently [73]. However, CD11b is also expressed on macrophages and neutrophils, making it impossible to distinguish these cellular populations in vivo [1]. Therefore, Yu et al. conjugated mouse CD11b and Gr-1 (also known as Ly6G) antibodies with different quantum dots that produce near-infrared (NIR)-IIa and NIR-IIb signals, respectively. Granulocytic MDSCs were identified in a noninvasive manner through colocalization of both fluorescence signals (Figure 3a) [46]. In a similar approach, we believe that monocytic MDSCs can be visualized using anti-CD11b and anti-Ly6C antibodies.

Cathepsin B and Z are produced mainly by tumor-infiltrating myeloid lineage cells. More than three-quarters of the cathepsin-inducible fluorescent probe ProSense 680 localizes to MDSCs and CD11b⁺F4/80⁺ macrophages [47]. However, cathepsin is also expressed on macrophages and thus cannot distinguish MDSCs from macrophages. Similarly, S100A8/A9 is mainly released by MDSCs [74, 75]. Eisenblaetter and colleagues assessed MDSC abundance in premetastatic lung tissue by SPECT/CT using an ¹¹¹In-labeled anti-S100A9 antibody. Their results indicated that the imaging signal in the premetastatic lung correlated significantly to the subsequent metastatic burden (Figure 3b) [48]. Considering the crucial role of MDSCs in priming metastatic niches, S100A8/A9-targeted imaging would be a potent probe to assess metastases [76].

However, because MDSCs share some cellular markers with TANs, further research is needed to find more specific markers and develop related probes.

2.4 Imaging of CAFs

FAP has emerged as the frontrunner to evaluate CAFs. Heidelberg et al. have developed quinoline-based FAP inhibitors (FAPIs), including FAPI-04 [49]. It shows rapid renal clearance and accumulation in tumors, leading to favorable tumor-to-background ratios (Figure 4) [49, 50]. Recently, a series of FAPI derivatives were also developed, such as FAPI-34 [78], FAPI-46 [51, 79], FAPI-74 [80], DOTA-2P(FAPI)₂ [52], ND-bisFAPI [53], and DOTA.SA.FAPi and its homodimer, DOTAGA.(SA.FAPi)₂ [81]. FAP-specific peptides, such as FAP-2286 and Alb-FAPtp-01, have longer tumor retention than the aforementioned small molecular FAPI derivates (Figure 4) [54, 82-84]. Considering that FAP-positive fibroblasts are present in many tumors, these tracers are promising pan-cancer probes, especially for tumors with low glucose metabolism [55]. Moreover, CAF-targeted radiopharmaceuticals have emerged as a novel and attractive strategy for cancer therapy, including those conjugated with ¹⁷⁷Lu, ²²⁵Ac, and ¹³¹I [84-87]. Additionally, a series of albumin binder-modified FAPI-derived radiotracers have been developed, such as EB-FAPI-B1 [88], TEFAPI-06, TEFAPI-07, FAPI-C12, and FAPI-C16 [89, 90]. With accumulating knowledge on immunosuppressive CAFs, we anticipate that these tracers will play a critical role in tumor therapy. However, FAP is frequently not overexpressed on all CAFs. For example, only one CAF subset (CAF-S1), but not three other subtypes (CAF-S2–CAF-S4), show upregulated FAP expression in human breast cancer [91]. Thus, developing probes targeting other CAF markers is necessary.

2.5 Imaging of T cells

Numerous reviews have discussed tracking T cells and imaging their antitumor immunity [92]. Therefore, we focus on immunosuppressive T cells and their biomarkers in the following section.

CD25 is a biomarker of regulatory T cells. The ligand of CD25, interleukin-2 (IL-2), has the potential to visualize Treg cells. ^99mTc-labeled IL-2 estimates T cell numbers correlated significantly to histological estimates of CD25⁺ T cells (r = 0.745, p = 0.001) [56]. Interestingly, IL-2 is present in many types of tumors [93]. Conjugating IL-2 with therapeutic nuclides might be a potential treatment.

Immune checkpoints suppress T cell-mediated antitumor activity. Therefore, the need for noninvasive appraisal of immune checkpoints is growing. A ⁶⁴Cu-labeled anti-mouse-PD-1 mAb (clone J43) specifically images PD-1 expression in tumors with a tumor-to-muscle uptake ratio as high as 11-fold (p < 0.05) [57]. ⁸⁹Zr-labeled nivolumab was developed to noninvasively image PD-1 expression in tumors and may allow assessment of inter- and intratumoral heterogeneity (Figure 5a) [58]. Similar probes include ⁸⁹Zr-Df-Pembrolizumab (Figure 5b) [59], ⁶⁴Cu-PD-1-Liposome-DOX [60], ⁶⁴Cu-NOTA-α-PD-1 (RMP1-14) [61], and ^99mTc-JS001 [62]. CTLA-4 is another immune checkpoint. An ⁶⁴Cu-anti-CTLA-4 mAb has displayed high accumulation in colon cancer models [63].

Together with advances in techniques such as single cell sequencing, exhausted CD8⁺ T cells are categorized into two major subsets: progenitor-exhausted T cells (Tex^prog) and terminally exhausted T cells (Tex^term) [94-96]. Accordingly, exhaustion markers such as T cell factor 1, lymphocyte activation gene 3, and T cell immunoglobulin and mucin domain-containing protein 3 may be useful to identify functionally different cytotoxic T cells. T cell visualization would facilitate identifying patients who might have a favorable response to immune checkpoint blockade.