Research ArticleRESEARCH METHODS

Chemokine receptor 4 targeted protein MRI contrast agent for early detection of liver metastases

See allHide authors and affiliations

Science Advances  07 Feb 2020:
Vol. 6, no. 6, eaav7504
DOI: 10.1126/sciadv.aav7504
  • Fig. 1 CXCR4 expression is up-regulated in UM cell lines, hepatic metastases in UM patients, and metastatic UM mice.

    (A) Tumor cells that express CXCR4 metastasize through CXCR4-CXCL12 interaction to specific organs that have intrinsically high concentrations of CXCL12 such as the lung, liver, and bone. (B) UM cell lines have elevated CXCR4 expression. Flow cytometry results measured elevated CXCR4 expression across different UM cell lines. Mel290 and M20-09-196 measured more than 80% of CXCR4 immunopositivity. Measurements of each cell line were done in triplicate. (C) CXCR4 IHC staining in liver tissue from metastatic UM patients (n = 4, IRS = 8.2 ± 1.3). The liver metastases displayed strong red intensity, denoting strong CXCR4 expression. (D and E) CXCR4 IHC staining of primary UM (D) and hepatic metastases (E) in metastatic UM mice. UM hepatic metastases have higher CXCR4 expression compared with primary UM, indicated by the red staining. (F) CXCR4 IRS of primary UM and metastases in the liver in metastatic UM mice. Hepatic UM metastases displayed stronger CXCR4 expression (IRS = 9.5 ± 0.8) than primary UM (IRS = 5.4 ± 0.3). P ≤ 0.05.

  • Fig. 2 ProCA32.CXCR4 binds to CXCR4.

    (A) Model structure of ProCA32.CXCR4 interacting with CXCR4 [Protein Data Bank (PDB): 4RWS] through targeting moiety. ProCA32.CXCR4 was constructed by engineering the CXCR4 targeting moiety (red) to ProCA32 (blue) by a flexible linker (green). The targeting moiety of ProCA32.CXCR4 binds to CXCR4 through residue-residue and electrostatic interactions. ProCA32.CXCR4 has two Gd3+ (red circle) binding sites. (B) CXCR4 targeting study of ProCA32.CXCR4 by ELISA. The dissociation constant of ProCA32.CXCR4 binding to CXCR4 was calculated as 1.10 ± 0.18 μM, measured by indirect ELISA. n = 3. The nontargeted contrast agent ProCA32 did not exhibit CXCR4 targeting capability. (C) Fluorescence staining of Mel290 cells to study the CXCR4 binding effect of ProCA32.CXCR4. Blue fluorescence is nucleus staining with 4′,6-diamidino-2-phenylindole (DAPI), green color is fluorescein-labeled ProCA32.CXCR4, red color indicates CXCR4 staining, and composite is the combination of nucleus, CXCR4, and ProCA32.CXCR4 staining. ProCA32.CXCR4 exhibited good spatial colocalization with CXCR4; Pearson’s r is 0.82. (D) Working flow of ProCA32.CXCR4. ProCA32.CXCR4 was administered through tail vein injection and distributed with blood flow, and specific targeting to CXCR4 high expression metastatic UM (indicated by black cells) was shown over time.

  • Fig. 3 Relaxivity (reported as “per Gd” value), transmetalation, and metal selectivity studies of ProCA32.CXCR4.

    (A) Relaxivity assessment of ProCA32.CXCR4 and GBCA with 60-MHz relaxometer; ProCA32.CXCR4 has 8 to 10 times higher r1 and r2 values than clinical GBCA. (B) Transmetalation study of ProCA32.CXCR4 and other GBCA in the presence of Zn2+. Thermodynamic index of ProCA32.CXCR4 incubated at 37°C with Zn2+ was 0.96, which is better than ProHance (0.93) and Gadovist (0.95). (C) Metal (Zn2+, Ca2+, Gd3+, and Tb3+) binding affinity and metal selectivity values of ProCA32.CXCR4 in comparison with clinical contrast agents. N/A, not available; PEG, polyethylene glycol.

  • Fig. 4 MRI images of metastatic UM mice M20-09-196 and histological correlation.

    (A) T1-weighted spin echo and T2-weighted fast spin echo MR images of M20-09-196 before and 48 hours after Cys-ProCA32.CXCR4 injection. At 48 hours after injection, both T1- and T2-weighted MR images revealed four lesions not observed before injection. The zoom-in view of the yellow rectangular region shows both gray and color scales. (B) H&E and IHC staining of M20-09-196 liver with UM metastases. H&E staining revealed four metastatic lesions, highlighted by different color circles, with similar locations as the metastases in MRI images. Higher-magnification images identified the growth pattern of metastases to be nodular pattern. S100 IHC labeling confirmed that the lesions were metastatic UM. CXCR4 immunohistological staining confirmed the CXCR4 expression on UM metastases. (C) The measurement of distances between metastases and the diameter of metastases in MRI images correlates with the H&E histological staining (y = 1.09x + 0.08). (D) Statistical analysis indicated that Cys-ProCA32.CXCR4 provides diagnostic validation for UM metastases in the liver. AUC = 0.84; P < 0.0001. Three mice were used for the experiment. Analyses were based on 11 metastases found on MR images. a.u., arbitrary units.

  • Fig. 5 Progressive MR images of the intrahepatic heterotopic xenotrasplantation UM mice (n = 3 for each group) with Cys-ProCA32.CXCR4 administration.

    (A) T1-weighted gradient echo MR images of control mice (with injection of nontargeted agent Lys-ProCA32) and mice with Cys-ProCA32.CXCR4 injection. MRI scans were acquired before and after injection at different time points until 46 hours; tumors are represented by the heat map in MRI images. (B) Percentage increase of SNR of melanoma tumors at different time points shows the dynamic binding process of Cys-ProCA32.CXCR4. For mice that received the Cys-ProCA32.CXCR4 injection, a gradual increase of intensity in melanoma tumor region was observed up to 24 hours, showing the CXCR4-targeting effect, followed by washing out at 46 hours (further time points not acquired). (C) Time plot of the liver SNR percentage increase following Cys-ProCA32.CXCR4 and Lys-ProCA32 injection. The liver SNRs of mice receiving Cys-ProCA32.CXCR4 and Lys-ProCA32 exhibited similar patterns of the SNR time plots, where the liver intensity substantially increased up to 3 hours after injection of both contrast agents, followed by loss of intensity after 3 hours. (D) Time plot of tumor rim and tumor center SNR change of mice with Cys-ProCA32.CXCR4 administration. Cys-ProCA32.CXCR4 exhibited good tumor permeability; tumor rim SNR was enhanced at early time points (12 min after injection). SNR enhancement gradually penetrated to the center of the tumor. At 24 hours after injection, the view of the tumor region following Cys-ProCA32.CXCR4 injection revealed broad distribution and heterogeneous enhancement. (E) Immunofluorescence staining of Cys-ProCA32.CXCR4 and Lys-ProCA32 in the liver (L) and tumor (T) of Mel290 mice. For mice that received Cys-ProCA32.CXCR4 injection (top), Cys-ProCA32.CXCR4 accumulated in the UM tumor tissue (denoted by red fluorescence). For the mice injected with Lys-ProCA32 (bottom), UM tumors exhibited dark fluorescence intensity relative to the UM tumor regions of the mice that received Cys-ProCA32.CXCR4 injection. (F) ICP-OES analysis of Gd3+ tissue distribution 2 days after injection of ProCAs. Mice with Cys-ProCA32.CXCR4 injection exhibited significantly more Gd3+ distribution in tumor tissue than mice that received Lys-ProCA32 injection (P < 0.01).

  • Fig. 6 Validating CXCR4 binding specificity by receptor blocking study (n = 3 for each group).

    (A) Comparison of subcutaneous UM tumor intensity change on T1-weighted MRI images following administration of Cys-ProCA32.CXCR4 with and without previous administration of blocking reagent; subcutaneous UM tumors are represented by color heat map. Tumor from UM mice that received Cys-ProCA32.CXCR4 injection showed significant increase in MRI signal intensity after Cys-ProCA32.CXCR4 administration. This enhancement could be blocked by first administrating the CXCR4 receptor blocking reagent. (B) Comparison of UM tumor SNR change following administration of Cys-ProCA32.CXCR4 and blocking reagent + Cys-ProCA32.CXCR4. For the mice that received Cys-ProCA32.CXCR4 injection, the SNR of UM tumor substantially increased at 24 hours after administration. This enhancement was blocked by first administrating a blocking reagent. As seen with the mice that received the blocking reagent and then the Cys-ProCA32.CXCR4 injection, the SNR of UM tumor was notably lower in comparison with the UM tumor SNR of the mice with Cys-ProCA32.CXCR4 administration. (C) UM tumor SNR change following administration of Cys-ProCA32.CXCR4, blocking reagent + Cys-ProCA32.CXCR4, and Lys-ProCA32. At 3 hours after administration, mice from all three groups showed an SNR increase. At 24 hours, mice with blocking reagent + Cys-ProCA32.CXCR4 administration and mice with Lys-ProCA32 administration showed SNR washout, while mice that received Cys-ProCA32.CXCR4 exhibited further SNR increases compared with 3 hours. At 48 hours, mice that received blocking reagent + Cys-ProCA32.CXCR4 and mice that received Lys-ProCA32 administration exhibited a further SNR decrease. Mice that received Cys-ProCA32.CXCR4 administration showed SNR washout at 48 hours in comparison with 24 hours.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/6/eaav7504/DC1

    Fig. S1. PEGylation SDS-PAGE gel of protein contrast agents.

    Fig. S2. Determination the relaxivity values of ProCA32.CXCR4.

    Fig. S3. Serum stability and transmetalation study of ProCA32.CXCR4.

    Fig. S4. Determination of ProCA32.CXCR4 metal binding affinities.

    Fig. S5. MRI images of metastatic UM mice M20-09-196 before and after administration of Lys-ProCA32, Eovist, and Lys-ProCA32.CXCR4 (n = 2 for Eovist group, n = 3 for Lys-ProCA32 and Lys-ProCA32.CXCR4 group).

    Fig. S6. T1-weighted MRI images of subcutaneous UM mice before and after administration of Cys-ProCA32.CXCR4, blocking reagent + Cys-ProCA32.CXCR4, and Lys-ProCA32 (n = 3 for each group).

    Fig. S7. Pharmacokinetic study of Cys-ProCA32.CXCR4 and ICP-OES analysis of Gd3+ content in different mouse organs.

    Fig. S8. H&E staining analysis of mice tissues collected 7 and 14 days after injection of Cys-ProCA32.CXCR4.

    Table S1. Relaxivities of investigated contrast agents in 10 mM Hepes at 37°C.

    Table S2. Clinical pathology profile of mouse serum.

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. PEGylation SDS-PAGE gel of protein contrast agents.
    • Fig. S2. Determination the relaxivity values of ProCA32.CXCR4.
    • Fig. S3. Serum stability and transmetalation study of ProCA32.CXCR4.
    • Fig. S4. Determination of ProCA32.CXCR4 metal binding affinities.
    • Fig. S5. MRI images of metastatic UM mice M20-09-196 before and after administration of Lys-ProCA32, Eovist, and Lys-ProCA32.CXCR4 (n = 2 for Eovist group, n = 3 for Lys-ProCA32 and Lys-ProCA32.CXCR4 group).
    • Fig. S6. T1-weighted MRI images of subcutaneous UM mice before and after administration of Cys-ProCA32.CXCR4, blocking reagent + Cys-ProCA32.CXCR4, and Lys-ProCA32 (n = 3 for each group).
    • Fig. S7. Pharmacokinetic study of Cys-ProCA32.CXCR4 and ICP-OES analysis of Gd3+ content in different mouse organs.
    • Fig. S8. H&E staining analysis of mice tissues collected 7 and 14 days after injection of Cys-ProCA32.CXCR4.
    • Table S1. Relaxivities of investigated contrast agents in 10 mM Hepes at 37°C.
    • Table S2. Clinical pathology profile of mouse serum.

    Download PDF

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article