IFNα gene/cell therapy curbs colorectal cancer colonization of the liver by acting on the hepatic microenvironment

Abstract Colorectal cancer (CRC) metastatic dissemination to the liver is one of the most life‐threatening malignancies in humans and represents the leading cause of CRC‐related mortality. Herein, we adopted a gene transfer strategy into mouse hematopoietic stem/progenitor cells to generate immune‐competent mice in which TEMs—a subset of Tie2+ monocytes/macrophages found at peritumoral sites—express interferon‐alpha (IFNα), a pleiotropic cytokine with anti‐tumor effects. Utilizing this strategy in mouse models of CRC liver metastasis, we show that TEMs accumulate in the proximity of hepatic metastatic areas and that TEM‐mediated delivery of IFNα inhibits tumor growth when administered prior to metastasis challenge as well as on established hepatic lesions, improving overall survival. Further analyses unveiled that local delivery of IFNα does not inhibit homing but limits the early phases of hepatic CRC cell expansion by acting on the radio‐resistant hepatic microenvironment. TEM‐mediated IFNα expression was not associated with systemic side effects, hematopoietic toxicity, or inability to respond to a virus challenge. Along with the notion that TEMs were detected in the proximity of CRC metastases in human livers, these results raise the possibility to employ similar gene/cell therapies as tumor site‐specific drug‐delivery strategies in patients with CRC.

Sham=mice transplanted with non-transduced HSPCs and intrasplenically injected with NaCl; CTRL=wild-type CB6 control mice. Note that at this low-dose infecting condition, LCMV viremia -which peaks by day 4 and disappears by day 7-8 post-infection when larger inocula are utilized experimentally (Iannacone et al, 2008) -remained undetectable throughout the course of the infection. No significant differences between the two groups were observed (by unpaired Student's t-Test). The decrease of HCT values observed in both Sham/CTRL and Tie2-IFNα mice represents a transient effect caused by the high frequency of blood withdrawal. Shortly after the end of the experiment the HCT values of both groups returned and remained to normal levels throughout the animals lifespan (as depicted in Fig EV2A). Data pooled from two independent experiments, Sham/CTRL (n=6), Tie2-IFNα (n=7); mean values are shown; error bars=S.E.M.   To identify the primary targets of Tie2-IFNα, reciprocal BM transplantations between IFNα/βR -/mice and syngeneic C57BL/6 mice were made (BM donor from which HSPCs were purified and recipient mouse strains are listed on the left). The resulting chimeric animals had restricted cellular compartments responsive to IFNα that could be analyzed (indicated on the right after the long arrows). HSPCs purification, transduction and transplantation were performed as previously described for CB6 mice. B. Quantification of vector copy number (VCN) performed by quantitative real-time PCR on white blood cells extracted from the chimeric mice described in A, 7 weeks posttransplant. n=number of transplanted mice per group; BM donor and recipient mouse strains are indicated at the bottom of panel C; data pooled from three independent experiments; mean values are shown; error bars=S.E.M. C. White blood cells (WBC, top panels), hematocrit (HCT, middle panels) and platelet counts (PLT, bottom panels) of the chimeric mice described in B. Note that as expected, both Tie2-GFP and Tie2-IFNα mice did not display significant differences (by unpaired Student's t-Test) in their basic hematological values even in C57BL/6 mice that received IFNα/βR -/transduced HSPCs despite the higher VCN, confirming the insensitivity of IFNα/βR -/-HSPCs to a cytokine such as IFNα that has potential to induce BM aplasia (Binder et al, 1997

Appendix Figure S5. Analysis of bone marrow reconstitution and immune cell activation in transplanted mice with established CRC liver metastases.
HSPCs purified from BM of syngeneic donor mice were enriched and transduced as previously described (see Appendix Supplementary Materials and Methods for details). Eight days before transplantation, 5x10 3 CT26 cells were injected under the liver capsule of anesthetized recipient mice. Twenty-eight days post-CT26 injection, recipient mice were bled and analyzed for BM reconstitution.

D, E. Flow cytometry characterization of immune CD4 + T cells (D) and CD8 + T cells (E) cells within splenocytes of Tie2-GFP mice (n=4) and
Tie2-IFNα mice (n=4) described above. Left panels depict absolute numbers (#) of total CD4 + and CD8 + T cells. Early activation markers such as CD25, CD69 and PD1 are used to characterize the number of activated cells within either CD4 + or CD8 + T cells in both groups of animals (center panels). Right panels show absolute number of the indicated subpopulations of cells within either CD4 + or CD8 + splenocytes. Naïve CD4 + or CD8 + T cell populations are defined as either CD4 + or CD8 + T cells which are also CD44 -/CD62L + ; central memory CD4 + or CD8 + T cell populations are defined as either CD4 + or CD8 + T cells which are also CD44 + /CD62L + ; effector CD4 + or CD8 + T cell populations are defined as either CD4 + or CD8 + T cells which are also CD44 + /CD62L -. Mean values and S.E.M. are shown; differences were not statistically significant by unpaired Student's t-Test.   The percentages depicted are calculated over the total CD45 + population in panels D and E; the mean fluorescent intensity (MFI) values of Tie2 + cells depicted in panels C and F were normalized by subtracting the background signal of the corresponding IgG isotype control.

Colorectal cancer cell lines
Cells were cultured under standard condition at 37°C in a humid atmosphere with 5% CO 2 in RPMI medium (Gibco) supplemented with 10% FBS (Lonza) and 1% penicillin/streptomycin/L-glutamin (Gibco). The H-2d-restricted CT26 colorectal cancer cell line (Brattain et al, 1980) was purchased from American Type Culture Collection (ATCC). The H-2b-restricted MC38 colorectal cancer cell line (Rosenberg et al, 1986) was kindly provided by P. wound clips for the skin as previously described (Kuruppu et al, 1996). In experiments utilizing IFNα/βR -/mice (inbred C57BL/6) or normal inbred C57BL/6 mice, only C57BL/6-derived MC38 cancer cells were used to avoid immune mediated rejection of mismatched cells.
In selected experiments, CRC cells were injected under the liver capsule as follow: i) mice were anesthetized by intraperitoneal injection of tribromoethanol (250mg/kg, Avertin); ii) through a midline incision, the left liver lobes were surgically exteriorized and 10µl of matrigel (BD Bioscience) containing 5x10 3 CT26 cells injected under the Glisson's capsule using a 29G needle as previously described (Kuo et al, 1995); iii) wounds were cauterized as described above (Kuruppu et al, 1996). Animals and biological samples were not anonymized for further analyses.

Plasmid constructs and LV production
Tie2-GFP and Tie2-IFNα_mirT plasmid constructs were generated as previously
Briefly, total BM cells were isolated from femurs and tibias of 7 to 10 week-old donor

Peripheral blood, intrahepatic leukocytes and splenocytes analyses
Seven to 10 weeks post-transplant, the whole anti-coagulated blood of Mock/Tie2-GFP and Tie2-IFNα mice was collected from the retro-orbital plexus of anesthetized animals (isoflurane, 5% for induction and 2% for maintenance in 2L/minute oxygen). Complete cell counts were measured in whole blood collected in 1/10 th volume of EDTA (45mg/ml; Sigma-Aldrich) utilizing an automated cell counter (HeCoVet, Seac-Radim). Circulating WBCs were obtained after red blood cell lysis with TQ-Prep workstation (Beckman-Coulter). Intrahepatic leukocytes (IHLs) were isolated from the liver of Tie2-GFP mice intrasplenically injected with NaCl or 5x10 3 CT26 at indicated time points as described previously (Sitia et al, 2011). In selected experiments, spleens of Tie2-GFP or Tie2-IFNα mice were removed at time of killing, smashed and filter through 40um nylon filter. transplanted mice were also analyzed for the expression of GFP in order to estimate the percentage of GFP + circulating monocytes, identified as: Cd11b + /Ly6C + /Ly6G -/GFP + cells. To assess VCN in Tie2-GFP and Tie2-IFNα mice, genomic DNA was isolated from about 3x10 5 WBCs/mouse using the Maxwell 16 automated extractor (Promega), quantified and analyzed by quantitative real-time PCR as previously described (Escobar et al, 2014).

Immunohistochemistry
At time of autopsy for each mouse, different organs were sampled, fixed in zincformalin, processed, embedded in paraffin, cut and stained with hematoxylin/eosin or further processed for immunohistochemical analyses as previously described (Guidotti et al, 1995;Sitia et al, 2011). Immunohistochemical staining was performed utilizing the following antibodies: anti-F4/80 (clone A3-1, AbD Serotec); anti-RFP (rabbit polyclonal, ab62341 AbCam); anti-CD34 (clone MEC14.7, Biolegend); anti-Ki67 (clone SP6, Neomarkers); anti-CD3 (clone SP7, AbCam); anti-CD45R/B220 (clone RA3-6B2, BD Pharmingen). All images were acquired using the Aperio Scanscope CS2 system (Leica Biosystems). To obtain the highest level of accuracy, the areas occupied by the RFP positive lesions in the liver were manually identified by three different operators blinded to any other information, summed and divided over the total liver area considered for analysis (between 12.4mm 2 and 44mm 2 per liver section), utilizing the ImageScope software (Leica Biosystems). All other quantifications were performed by automated image analysis software through dedicated macros of the ImageScope program, customized following manufacturer's instructions (Leica Biosystems). The images shown were identified as representative area of interest within the total area of the specimen analyzed and exported as ImageScope snapshots. Image processing was performed post-analysis by Phostoshop CS4 (Adobe Systems Software). BioImaging Center (ALEMBIC). 15-18µm z-stacks were projected in 2D and processed using Fiji image processing software (Schindelin et al, 2012) and Photoshop CS4

Immunofluorescence and confocal microscopy
(Adobe Systems Software).

RNA extraction and quantitative RT-PCR gene expression analyses
Total RNA was isolated from liver homogenates by phenol-chloroform extraction as previously described (Guidotti et al, 1995). The extracted RNA was subsequently retro-

LCMV infection and related procedures
The Armstrong strain of LCMV was utilized (Iannacone et al, 2008). Eighty-four or 54 days after receiving saline or CT26 CRC cells, Mock-transduced or wt mice (CTRL) or Tie2-IFNα mice were intraperitoneally infected with 200 pfu of LCMV. Whole blood was collected at the indicated time points from the retro-orbital plexus of anesthetized mice and white blood cell, hematocrit and platelet values were measured with an automated cell counter (HeCoVet, Seac-Radim). Single-cell suspensions were prepared from whole blood harvested at day 8 post-infection as previously described (Iannacone et al, 2008). Cells were stained with Pacific Blue-conjugated anti-CD8 (clone 53-6.7; BD Pharmingen) and allophycocyanin-conjugated anti-IFNγ (clone XMG1.2; BD Pharmingen). LCMV-specific CD8 + T cells were identified by immunostaining with Pacific Blue-conjugated anti-CD8 (clone 53-6.7; BD Pharmingen) and by Phycoerythrinconjugated recombinant soluble dimeric H-2d/Ig fusion protein (BD Pharmingen) complexed with the immune-dominant H-2d-restricted LCMV NP 118-126 peptide (Isogawa et al, 2005). Samples were analyzed by flow cytometry with FACS CantoII (BD Pharmingen) and data were processed using FlowJo software (Tree Star Inc.). The presence of LCMV genomes in serum collected at the same time points abovementioned was quantified by real-time PCR using the following LCMV-specific primers 5'-CTCCTTTCCCAAGAGAAGACTAAG-3' and 5'-TCCATTTGGTCAGGCAATAAC-3' as previously described (Iannacone et al, 2008). All infections were performed in designated BSL-2 or BSL-3 workspaces, in accordance with institutional guidelines.