NK Cells Eradicate Gal-1 Deficient Glioma

NK Cells Eradicate Gal-1 Deficient Glioma

Supplementary Figure Legends

Supplementary Figure S1. Plasmid map of the pCI-neo-mCitrine expression vector.

The human cytomegalovirus (hCMV) promoter provides constitutive expression of mCitrine fluorescent protein. A neomycin-resistance cassette is included to select for transgene expression.

Supplementary Figure S2. Gal-1 expression is positively correlated with histological grade in human glioma.

(A) Patient survival data from the REpository for Molecular BRAin Neoplasia DaTa (REMBRANDT) database stratified by gal-1 expression among 343 patients diagnosed with glioma. Gliomas with upregulated gal-1 expression (red curve) carry worse prognosis compared to those with intermediate (yellow curve; p<0.00000001 vs. upregulated) and downregulated (green curve; p<0.00000001 vs. upregulated) gal-1 expression. (B) Table comparing LGALS1 expression from the 343 patients represented in panel A stratified by histological grade. As LGALS1 expression increases the proportion of patients carrying grade IV gliomas increases.

Supplementary Figure S3. Validation of GL26-Cit cells independently transduced with gal-1-specific or control shRNA.

(A) Western blot for gal-1 in GL26-Cit cell-lines demonstrating the lack of gal-1 knockdown in each control cell-line (i.e. EV1, EV2, NT1, NT2; independently transduced replicates) compared to efficient knockdown in three independent transduction experiments where single-cell GL26-Cit-gal1i clones were isolated (i.e. C01; -84.85% knockdown vs. NT, C02; -76.52% knockdown vs. NT, and C03; -79.55% knockdown vs. NT). (B) GL26-Cit-gal1i clones (C01, C02, C03) implanted into the striatum of RAG1-/- mice. Like the mixed GL26-Cit-gal1i population, each clone also fails to diffusely invade the brain after 48hrs in-vivo and is characterized by a column of tumor cells that fail to leave the initial site of tumor implantation.

Supplementary Figure S4. GL26-Cit-gal1i cells do not undergo spontaneous cell-death in-vitro.

GL26-Cit-gal1i and GL26-Cit-NT cells were initially seeded at 1x105 cells/T25 tissue culture flask (3 independent flasks per group), which were sequentially harvested and analyzed by Typan blue dye exclusion every 24hrs for 3-days.

Supplementary Figure S5. Morphometric analysis of 48hr post-implantation intracranial GL26-Cit gliomas (shown in Fig. 1G).

Glioma morphologies were characterized as being either compact (gal-1-), or diffuse (gal1+), as illustrated by the cartoons to the left and right of the radar chart, respectively. Total tumor area was subdivided into color-coded cell-clusters based on size, and graphed on the radar chart shown at center. Color-coded lines on the chart represent cell-clusters of various sizes as indicated by the key below. Radial black arrows correspond to a fluorescence confocal micrograph of the indicated glioma. The point at which each color-coded line intersects the black arrow specifies the percentage of total tumor area made up of cell-clusters of that size. Cell-clusters of each glioma sum to 100%. The four GL26-Cit-gal1i tumors (left-hand side of chart) display a skewed distribution towards large cell-clusters 104 pixels in size (blue color-coded line) (***p<0.001; 66.10 ± 3.73% in gal-1- tumors vs. 47.10 ± 3.188% in gal-1+ tumors; two-way ANOVA followed by Tukey’s post-test). Conversely, the five GL26-Cit tumors (right-hand side of chart) exhibit a significantly higher proportion of total tumor area made up of smaller cell clusters 104 - 103 pixels in size (red color-coded line) (***p<0.001; 17.86 ± 4.07% in gal-1+ tumors vs. 3.42 ± 2.18% in gal-1- tumors; two-way ANOVA followed by Tukey’s post-test).

Supplementary Figure S6. Validating immunodepletion of NK cells and basophils in RAG1-/- mice.

(A) Flow cytometric analysis with antibodies against NK1.1 and CD3ε, makers of NK cells and T-cells, respectively, on whole splenocytes harvested from RAG1-/- mice treated with two doses of normal rabbit serum (left dot plot) or anti-asialo GM1 (right dot plot) each separated by 3-days (n=6; 3 mice/treatment group). Mice were euthanized 5-days after the initiation of treatment. The average percentage of NK1.1+/CD3ε- cells (i.e. NK cells) present within each treatment group as shown. (B) Treatment with anti-asialo GM1 led to a ~86% reduction in splenic-derived NK cells compared to mice treated with normal rabbit serum after two doses of anti-asialo GM1 (*p=0.0356; 8.42 ± 2.26% normal rabbit serum vs. 1.20 ± 0.49% anti-asialo GM1; unpaired, two-tailed, Student’s t-test), thus validating anti-asialo GM1 as an in-vivo NK depletion agent in our model. (C) Basophil depletion with anti-CD200R3 monoclonal antibodies in RAG1-/- mice bearing GL26-Cit-gal1i cells after 10dpi. Top epifluorescence show the brains of 3 mice treated with rat IgG2b isotype control antibodies. Bottom epifluorescence show the brains of 3 mice treated with anti-CD200R3 (n=6; 3mice/treatment group). Both agents failed to permit gal-1 deficient glioma growth in RAG1-/- mice.

Supplementary Figure S7. Anti-CD3ε immunolabeled brain tissue sections from mouse A1 (top image) and mouse N2 (bottom image) shown in Fig. 5I.

Numerous tumor-infiltrating CD3ε+ cells can be seen in the brains of mice treated with ASGM1, while NRS treated mice lack these cells.

Supplementary Figure S8. Splenic-derived RAG1-/- NK cell purity assessment.

Representative histogram of post-sort purity assessment of NK1.1+/CD3ε- FACS sorted RAG1-/- splenocytes used for in-vitro NK-mediated cytotoxcity assays. Isolated NK cells were consistently >95% pure and >99% viable by Trypan dye exclusion.

4