Glycoengineering of chimeric antigen receptor (CAR) T-cellsto enforce E-selectin bindingReceived for publication, September 18, 2019, and in revised form, October 16, 2019 Published, Papers in Press, October 18, 2019, DOI 10.1074/jbc.RA119.011134

X Nandini Mondal‡§1, Mariana Silva‡, Ana P. Castano¶, Marcela V. Maus¶, and Robert Sackstein‡§�**2

From the ‡Department of Dermatology and Harvard Skin Disease Research Center, and the §Program of Excellence inGlycosciences, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115, the �Department ofMedicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115, the ¶Cellular ImmunotherapyProgram, Cancer Center, Massachusetts General Hospital, Boston, Massachusetts 02129, and the **Department of TranslationalMedicine, Herbert Wertheim College of Medicine, and Translational Glycobiology Institute, Florida International University, Miami,Florida 33199

Edited by Gerald W. Hart

Tissue colonization (homing) by blood-borne cells criticallyhinges on the ability of the cells to adhere to vascular endothe-lium with sufficient strength to overcome prevailing hemody-namic shear stress. These adhesive interactions are most effec-tively engendered via binding of the endothelial lectin E-selectin(CD62E) to its cognate ligand, sialyl Lewis-X (sLeX), displayedon circulating cells. Although chimeric antigen receptor (CAR)T-cell immunotherapy holds promise for treatment of varioushematologic and non-hematologic malignancies, there is essen-tially no information regarding the efficiency of CAR T-cellhoming. Accordingly, we performed integrated biochemicalstudies and adhesion assays to examine the capacity of humanCAR T-cells to engage E-selectin. Our data indicate that CART-cells do not express sLeX and do not bind E-selectin. However,enforced sLeX display can be achieved on human CAR T-cells bysurface fucosylation, with resultant robust E-selectin bindingunder hemodynamic shear. Importantly, following intravascu-lar administration into mice, fucosylated human CAR-T cellsinfiltrate marrow with 10-fold higher efficiency than do unfuco-sylated cells. Collectively, these findings indicate that custominstallation of sLeX programs tissue colonization of vascularlyadministered human CAR T-cells, offering a readily translatablestrategy to augment tissue delivery, thereby lowering the perti-nent cell dosing and attendant cell production burden, for CART-cell immunotherapy applications.

Immunotherapy is an emerging approach for cancer treat-ment that offers the opportunity to specifically target tumor

cells and thereby avoid systemic toxicities of conventionalradiochemotherapy. One major advance in this approach is chi-meric antigen receptor (CAR)3 T-cell therapy, which has showncurative potential against B-cell malignancies and holds greatpromise for treating other hematologic malignancies and solidtumors. However, general applicability of this therapeuticapproach is limited by production costs related to the culture-expansion of requisite numbers of CAR T-cells to achievetherapeutic response(s). Furthermore, studies indicate thatsystemic toxicities result from dose-related off-target dissemi-nation of these cytotoxic lymphocytes (1–3). Thus, to optimizethe applicability of CAR T-cell therapeutics, it is critical todevise strategies to improve lesion-specific colonization of this“live drug,” thereby improving efficacy and dropping produc-tion costs and incidence of adverse events.

It is well known that tumor regression by intravascularlyadministered CAR T-cells depends on the ability of the cells toinfiltrate affected site(s), whether within bone marrow paren-chyma for hematologic malignancies, or within primary solidtumor beds, or within metastatic tumor foci (4, 5). As opposedto non-specific entrapment, which results in stochastic cellularaccumulation in both unaffected and lesional tissue, the abilityof circulating cells to engage vascular endothelium allows for“homing” in which immunotherapeutic cells selectively colo-nize target sites. Homing is initiated by tethering and rolling ofblood-borne cells onto target endothelium (6), a process mostefficiently mediated by E-selectin receptor–ligand interactions.E-selectin is an endothelial lectin that is constitutively ex-pressed by bone marrow and skin microvessels (7), is potentlyinduced within essentially all microvascular endothelial beds byinflammatory cytokines (e.g. TNF� and interleukin-1�) (8), andis also characteristically expressed by cancer microvessels (9).Notably, bone is a common metastastic site for a variety of solidmalignancies, and a recent study reported that marrowmicrovessel expression of E-selectin promotes bone metastasis

This work was supported by National Institutes of Health Grant PO1-HL107146 and the Team Jobie Fund (to R. S.). According to National Insti-tutes of Health policies and procedures, the Brigham and Women’s Hospi-tal has assigned intellectual property rights regarding cell surface glycanengineering to R. S., and R. S. has licensed portions of this technology to anentity he has founded (Warrior Therapeutics, LLC), to BioTechne, Inc., andto Mesoblast Ltd. R. S.’s ownership interests were reviewed and are man-aged by Brigham and Women’s Hospital and Partners HealthCare inaccordance with their conflict of interest policy.

This article contains Figs. S1–S3 and Table S1.1 Present address: Beth Israel Deaconess Medical Center, 330 Brookline Ave.,

Boston, MA 02215.2 To whom correspondence should be addressed: Herbert Wertheim College

of Medicine, Florida International University, 1200 SW 8th St., AHC2-693,Miami, FL 33199. Tel.: 305-348-0618; Fax: 305-348-0123; E-mail:sackstein@fiu.edu.

3 The abbreviations used are: CAR, chimeric antigen receptor; TNF, tumornecrosis factor; EGFR, epidermal growth factor receptor; sLeX, sialyl Lew-is-X; HS, human AB serum; NT, nontransduced; Gal, galactose; Neu5Ac,N-acetylneuraminic acid; Fuc, fucose; HUVEC, human umbilical vein endo-thelial cell; BT, buffer-treated; TBS, Tris-buffered saline; ANOVA, analysis ofvariance; BM, bone marrow; LacNAc, lactosamine; GlcNAc, N-acetylgluco-samine; IL-2, Interleukin-2.

croARTICLE

J. Biol. Chem. (2019) 294(48) 18465–18474 18465© 2019 Mondal et al. Published under exclusive license by The American Society for Biochemistry and Molecular Biology, Inc.

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

at B

iomedical L

ibrary, UC

SD on January 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

at B

iomedical L

ibrary, UC

SD on January 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

at B

iomedical L

ibrary, UC

SD on January 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

of cancer cells (10). Therefore, the ability of CAR T-cells tohome to E-selectin– bearing sites such as marrow is critical forprecise targeting of osteotropic metastatic cancers such asprostate, breast, and lung adenocarcinomas, as well as forhematologic malignancies such as acute leukemias and multi-ple myeloma.

The tetrasaccharide glycan known as “sialyl Lewis X” (sLeX,CD15s) is the canonical binding determinant of E-selectin (6).sLeX is a sialo-fucosylated lactosaminyl glycan, displayed at thetermini of specialized membrane glycoproteins (11) and glyco-lipids (12) of leukocytes. Although sLeX expression on circulat-ing native human T-cells is well-characterized, no prior studyhas evaluated the expression of sLeX by human CAR T-cells.Indeed, to date, there is no information regarding the ability ofCAR T-cells to engage endothelial cells under hemodynamicflow conditions. Here, using CAR T-cells expressing antibodyspecificity for human epidermal growth factor receptor(EGFR), a clinically targetable cell membrane protein highlyamplified in many types of cancer (13, 14), we report that typicalin vitro conditions used for CAR T-cell propagation/expansiondeaden cell surface sLeX display, leading to a commensuratereduction in E-selectin–mediated tethering and rolling onendothelial cells under shear stress conditions. However, cul-ture-expanded CAR T-cells display uniformly high levels oftype 2 sialyllactosamines (sialylLacNAc) that can be convertedto sLeX via enzyme-based cell surface fucosylation (exofucosy-lation) (6). This enforced sLeX expression yields significantlyhigher CAR T-cell tethering and rolling adhesive interactionson endothelial cells expressing E-selectin, and, upon intravas-cular injection into mice, these cells infiltrate bone marrowwith �10-fold higher efficiency than unfucosylated CART-cells. Collectively, these findings indicate that deficits in CART-cell homing can be remedied by cell surface glycoengineer-ing, providing a readily translatable strategy for improving col-onization of CAR T-cells within marrow and other tissueswhose endothelial beds express E-selectin.

Results and discussion

Human CAR T-cells directed against human EGFR, which ishighly amplified in various cancers (13, 14), were manufacturedby lentiviral transduction of purified human T-cells with thehuEGFR-BBZ CAR construct co-expressing mCherry to reporttransduction. T-cells were stimulated with anti-CD3/CD28microbeads before transduction with CAR construct and cul-ture-expanded for 10 days in growth medium supplementedwith either FBS or human AB serum (HS) and IL-2. The 10-dayexpanded human CAR T-cells were then co-cultured with U87cells (an EGFR-expressing human cell line) for 7 additional daysto allow antigen-specific expansion (15).

We first sought to assess whether culture expansion modifiessLeX display on CAR T-cells (Fig. 1). To this end, we measuredbinding of the mAb HECA452 (which recognizes sLeX) to CART-cells manufactured as described above. We observed thatnative (i.e. freshly obtained from normal blood) human T-cellsexhibit heterogeneity in sLeX expression, with an average of�25% of CD4� and �15% of CD8� T-cells characteristicallyexpressing sLeX (Fig. 1B) (11). However, culture-expansion inmedium containing either FBS (Fig. 1, C–F) or HS (Fig. 2B, right

panel) leads to a significant reduction in CAR T-cell sLeX

expression compared to that of native T cells. To determinewhether transduction by CAR construct itself alters sLeX dis-play, and to evaluate whether culture-expansion differentiallyaffects CD4� or CD8� T-cells, we divided the CAR-transducedand expanded T-cells into the following subpopulations basedon mCherry expression and CD4 staining (Fig. 1C): CD4�CART-cells (CD4�mCherry�), CD4�nontransduced (NT) T-cells(CD4�mCherry�), CD8�CAR T-cells (CD8�mCherry�), andCD8� NT T-cells (CD8�mCherry�). After 10 days in culture,�3– 8% of CD4� and �8%–11% of the CD8� T-cells (bothCAR and NT T-cells) express sLeX (Fig. 1D). At this time, thesurface density of sLeX on the minor population of sLeX� cellswas also low. Co-culturing these cells for 7 additional days withU87 cell line further reduces sLeX expression (Fig. 1F); after 17

Figure 1. Culture expansion progressively depletes sLeX expression onCAR T-cells, whereas transduction with the CAR construct has no effecton sLeX display. A–F, sLeX expression (measured by mAb HECA452 binding)on the surface of native human T-cells (A and B), 10-day expanded (C and D),and 17-day expanded (E and F) CAR-transduced T-cells. A, C, and E: represen-tative flow cytometry contour plots presenting mCherry expression and CD4staining of native (A), 10-day (C), and 17-day (E) expanded CAR-transducedT-cells. mCherry� cells indicate CAR T-cells and mCherry� cells indicate NTT-cells. Numbers within each gate represent percentage of positive cells. B, D,and F: representative flow cytometry histograms presenting sLeX expressionon native CD8� (B, left) and CD4� (B, right) T-cells; 10-day CD8� CAR T-cells (D,top left) and CD4� CAR T-cells (D, top right); CD8� NT T-cells (D, bottom left)and CD4� NT T-cells (D, bottom right); 17-day CD8� CAR T-cells (F, top left) andCD4� CAR T-cells (F, top right); and CD8� NT T-cells (F, bottom left) and CD4�

NT T-cells (F, bottom right) . Numbers above each gate represent the percent-age of positive cells. Mean � S.D. of three to four experiments.

Site-directed delivery of CAR T-cells

18466 J. Biol. Chem. (2019) 294(48) 18465–18474

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

days, on an average, only �3% of cells express sLeX in bothCD4� and CD8� T-cell compartments (CAR or NT T-cells),and these cells display very low sLeX surface density. Notably,upon expansion, both CAR and NT T-cells drop sLeX levels bysimilar proportions. Together, our results indicate that culture-expansion progressively deadens expression of the tetrasaccha-ride sLeX within both CD4� and CD8� T-cell compartments,

and, importantly, transduction by lentivirus encoding CAR con-struct in itself has no effect on sLeX display. In other words,culture-expansion itself, in either FBS or HS, markedly damp-ens sLeX display.

The tetrasaccharide sLeX (Neu5Ac-�(2,3)-Gal-�(1,4)-[Fuc-�(1,3)-]GlcNAc�1-R) is composed of a type 2 lactosamine(type 2 LacNAc) disaccharide core in which a galactose (Gal)

Figure 2. Exofucosylation custom-installs sLeX display on culture-expanded CAR T-cells and augments CAR T-cell E-selectin binding function. A,schematic of the glycosyltransferase-programmed stereosubstitution (exofucosylation) reaction, in which type 2 sialylLacNAc acceptors present on the cellsurface are converted to sLeX using purified �(1,3)-fucosyltransferases (FT6 or FT7) and GDP-fucose. B i, sLeX expression (measured by HECA452 binding) onnative human T-cells that were either treated with buffer alone (BT, black line) or exofucosylated with purified FT6 enzyme (FT6, red line). B ii, sLeX expression onUT or CAR T-cells expanded using medium containing 10% HS and IL-2 for10 days: Left, representative histograms presenting sLeX expression as measured byHECA452 binding (Black lines represent BT, and red lines represent FT6); Right, aggregate data presenting the mean fluorescence intensity (MFI) of HECA452binding to 10-day UT and CAR T-cells with or without exofucosylation by FT6. n � 3 (each data point represents an independent T-cell donor). Ratio paired t testcomparing BT and FT6 treatment groups: *, p � 0.05; **, p � 0.01. C and D: Culture-expanded T-cells that had undergone transduction were either treated withbuffer alone (BT) or exofucosylated with FT6 (FT6), followed by staining with mAbs against CD4 and sLeX (HECA452). mCherry� indicates CAR T-cells that weresuccessfully transduced and mCherry� indicates NT T-cells. The bar plots represent mean fluorescence intensity (MFI) of HECA45 binding. C, sLeX expression byBT (gray columns) and FT6-treated (red columns) CD4� T-cell populations; Left to right: native, 10-day NT, 10-day CAR T, 17-day NT, and 17-day CAR T-cells. n �3– 4. Error bars represent S.E.; Ratio paired t test: *, p � 0.05; **, p � 0.01; ***, p � 0.001. D, sLeX expression on BT (gray columns) and FT6-treated (red columns)CD8� T-cell populations; Left to right: native, 10-day NT, 10-day CAR, 17-day NT, and 17-day CAR T-cells. n � 3– 4. Error bars indicate S.E.; Ratio paired t test: *, p �0.05; **, p � 0.01; ***, p � 0.001. E, quantitative RT-PCR analysis of the human �(1,3)-fucosyltransferase enzyme transcripts FUT3, FUT5, FUT6, and FUT7 in 10- and17-day-expanded CD4� (top panel) and CD8� (bottom panel) T-cell populations. Gene expression values are presented as log2 -fold change compared withexpression of each gene in native (freshly isolated) human CD4� or CD8� T-cells. n � 2 independent donors, triplicates. Data are presented as mean � S.E.

Site-directed delivery of CAR T-cells

J. Biol. Chem. (2019) 294(48) 18465–18474 18467

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

is �(1,4)-linked to N-acetylglucosamine (GlcNAc) (i.e.Gal-�(1,4)–GlcNAc). When this LacNAc is modified by a sialicacid (N-acetylneuraminic acid, Neu5Ac) residue in �(2,3)-link-age to Gal, this structure is called a “type 2 sialylLacNAc”(Neu5Ac-�(2,3)-Gal-�(1,4)-GlcNAc�1-R). Addition of afucose (Fuc) residue in �(1,3)-linkage to the GlcNAc withintype 2 sialylLacNAc creates sLeX. All of these reactions are dic-tated by glycosyltransferases, which function in an assemblyline–like sequence. As such, type 2 sialylLacNAc is the precur-sor of sLeX, and, when present on the cell surface, this trisac-charide acceptor can be converted to sLeX by glycosyltrans-ferase-programmed stereosubstitution, wherein fucose isinstalled at �(1,3)-linkage to the GlcNAc in vitro using purified�(1,3)-fucosyltransferase enzyme (fucosyltransferase 6 (FT6))and the nucleotide sugar donor GDP-fucose (Fig. 2A) (6).Accordingly, we sought to assess whether CAR T-cells culture-expanded according to conventional methods express type 2sialylLacNAc that can be converted to sLeX by exofucosylation.We observed that exofucosylation enforces sLeX display onT-cells expanded in medium containing HS (Fig. 2B, ii) or FBS(Fig. 2, C and D). Notably, even after exofucosylation, only afraction of the native (i.e. unexpanded) T-cells display sLeX

(45% of CD4� and 35% of CD8� T-cells) (Fig. 2B, i, and Fig.S1A), indicating that a significant proportion of native T-cellsdo not innately express type 2 sialylLacNAc. Strikingly, all fouraforementioned culture-expanded T-cell subsets, as identifiedbased on mCherry expression and CD4 staining (Fig. 1C) (i.e.CD4� or CD8� CAR or NT T-cells), display uniformly highlevels of sLeX after exofucosylation (Fig. 2, C and D, and Fig. S1,B and C). Furthermore, untransduced (UT) T-cells (nevertransduced but expanded under similar conditions as CAR-transduced cells) exhibit similar boost in sLeX display uponexofucosylation, as do NT or CAR T-cells (Fig. 2B, ii). Impor-tantly, post-exofucosylation, similar sLeX levels were observedon CAR and NT T-cells, indicating that procedures/techniquesassociated with CAR transduction do not alter the expression oftype 2 sialylLacNAc acceptors.

Our finding that exofucosylation is able to complete sLeX

tetrasaccharide structure on the surface of CAR T-cells, and isable to restore E-selectin binding function of CAR T-cells, indi-cates that there are sialylated acceptors present on the rightprotein backbones on CAR T-cell surface, but the innate fuco-sylation machinery is down-regulated in these cells. We testedthis hypothesis by performing quantitative RT-PCR of native(freshly isolated) human T-cells, and, NT and CAR T-cells cul-tured for 10 or 17 days to assess expression of �(1,3)-fucosyl-transferase (FUT) enzymes that add fucose to the type 2 sialyl-LacNAc acceptors (FUT3, FUT5, FUT6, and FUT7). We foundthat each of the �(1,3)-FUT enzymes is significantly down-reg-ulated in CD4� and CD8� NT or CAR T-cells (Fig. 2E).

To evaluate the E-selectin–reactive glycoproteins engen-dered by exofucosylation of native and culture-expandedT-cells, we performed Western blot analysis using E-selectin–Ig chimera as a probe. Native T-cells show relativelymodest E-selectin–Ig reactivity at a band of 150-kDa molecularmass, which comprises two glycoprotein E-selectin ligands:The PSGL1 glycoform known as cutaneous lymphocyte antigen(CLA) and the CD43 glycoform that binds E-selectin (CD43E)

(11). A higher-molecular-mass-band (�250 kDa) is also found,representing the homodimer of PSGL1. Upon exofucosylationof native T cells, the E-selectin binding capacity of each of theseglycoproteins increases dramatically (Fig. 3A). In contrast, cul-ture-expanded T-cells, both CAR and NT T-cells, do notexhibit any appreciable E-selectin binding by Western blotanalysis. However, upon exofucosylation, markedly increasedE-selectin binding was evident on CAR T-cells culture-ex-panded for 10 days (Fig. 3A) and for 17 days (Fig. S1E). Immu-noprecipitation by E-selectin–Ig followed by Western blottingindicates that exofucosylation profoundly increases both cuta-neous lymphocyte antigen and CD43E expression on CART-cells, and also enforces expression of the E-selectin– bindingglycoform of CD44 known as Hematopoietic Cell E-/L-selectinLigand (HCELL) (Fig. 3B) (16). Together, these results indicatethat exofucosylation can install high-level sLeX display on themembrane glycoproteins PSGL-1, CD43, and CD44 of culture-expanded CAR T-cells.

To investigate whether installation of sLeX on CAR T-cells byexofucosylation translates to functional E-selectin–mediatedtethering and rolling activity, we performed in vitro microflu-idic flow chamber assays to evaluate T-cell–adhesive interac-tions on E-selectin– expressing endothelial beds under hemo-dynamic shear conditions. To this end, we perfused untreatedor exofucosylated T-cells over TNF�-stimulated HUVECmonolayers at 2 dynes/cm2 shear stress (Fig. 4A). These condi-tions closely mimic physiological environment prevailingwithin microvasculature (17). As expected, based on a priorstudy (11), native T-cells display robust rolling on E-selectin–expressing HUVEC monolayers; however, we observed only amoderate increase in the rolling flux of native T-cells after exo-fucosylation (Fig. 4C). Notably, consistent with diminished cellsurface sLeX levels, both 10- and 17-day– expanded T-cells(CAR and NT) display profoundly impaired tethering and roll-ing. Importantly, exofucosylation leads to strikingly higher roll-ing and adhesion of culture-expanded T-cells (both CAR andNT) on E-selectin– bearing HUVEC monolayers (Fig. 4, B andC). These findings indicate that conventional culture tech-niques markedly induce expression of type 2 sialylLacNAc gly-cans on the T-cell surface. Notably, transduction with CARconstruct does not itself modify expression of sialylLacNAcstructures, as NT and CAR T-cells each exhibit similar rollingand adhesion flux upon exofucosylation.

Figure 3. Identification of the proteins carrying sLeX in exofucosylatedCAR T-cells. A, Western blot using E-selectin–Ig chimera as a probe of BT- orFT6-treated native T-cells, 10-day CAR T-cells, and 10-day NT T-cells. GAPDHstaining serves as a loading control. B, Western blot analysis of E-selectin–reactive glycoproteins of exofucosylated CAR T-cells. E-selectin–Ig immuno-precipitated glycoproteins from FT6-exofucosylated CAR T-cells were stainedwith antibodies against PSGL1 (left), and CD43 and CD44 (right). The numberson the left indicate molecular mass in kilodaltons.

Site-directed delivery of CAR T-cells

18468 J. Biol. Chem. (2019) 294(48) 18465–18474

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

Given that enforcing sLeX display on CAR T-cells enhancestheir adhesive interactions on E-selectin– bearing endothelialcells in vitro, we reasoned that exofucosylation may alsoimprove the ability of CAR T-cells to home to tissue sites wheremicrovascular endothelial cells express E-selectin. Since bonemarrow microvessels constitutively express E-selectin (7), weused homing to bone marrow as a model to test in vivo tissue-specific homing of exofucosylated CAR T-cells. Importantly,many solid malignancies that amplify membrane EGFR expres-sion are osteotropic, providing key rationale for analyzing thecapacity of EGFR-directed CAR T-cells to colonize bone mar-row. Previous studies from our laboratory that have directlyassessed (via real-time intravital microscopy) the capacity ofsystemically administered human cells to extravasate (includ-

ing hematopoietic stem cells (18), mesenchymal stem cells (16,19 –21), iPSC-derived hematopoietic stem cells (22), and leu-kocytes (18, 23)), have consistently revealed that enforcedexpression of E-selectin ligands uniformly leads to transen-dothelial migration at endothelial beds that express E-selec-tin. Thus, an increased human cell count in the marrow is adefinite measure of marrow parenchymal infiltration.Accordingly, we intravenously injected varying numbers ofbuffer-treated (BT) and FT6-exofucosylated (FT6) CART-cells into immunocompromised mice and collectedflushed marrow cells to quantify parenchymal infiltrates ofhuman T-cells (Fig. 5). We administered five cell doses foreach treatment (BT or FT6) and collected marrow cells 24 hpost-injection (Fig. 5B).

Figure 4. Exofucosylation augments tethering and rolling function of culture-expanded CAR T-cells on TNF�-stimulated HUVECs. A, schematic of theexperiment evaluating ex vivo tethering and rolling function of CAR-transduced and culture-expanded T-cells. T-cells were either treated with buffer alone (BT)or exofucosylated with FT6. The 10- and 17-day-expanded CAR-transduced T-cells were then flow-sorted based on mCherry expression to isolate NT (i.e.mCherry�) and CAR T-cells (i.e. mCherry� cells). These cells were then perfused over a monolayer of TNF�-stimulated HUVECs at 2 dynes/cm2 shear stress at 2 106/ml cell concentration. Rolling and adherent cells were counted and normalized to cell count per unit area. B, quantitation of E-selectin–mediatedrecruitment of BT (black columns) and FT6-treated (red columns) T-cell populations on TNF�-stimulated HUVECs. The plots present normalized counts of rolling(solid columns) and firmly adherent (open columns) T-cells. No T-cell rolling was observed in the absence of HUVEC stimulation. n � 2–3; ordinary one-wayANOVA (p � 0.01) with Sidak’s multiple comparison test comparing BT with FT6-treated cells for each T-cell population: *, p � 0.05; **, p � 0.01. C, Column plotpresenting normalized counts of rolling (solid columns) and firmly adherent (open columns) cells on HUVEC monolayers for native T-cells (nontransduced andnonexpanded, freshly isolated), NT T-cells (nontransduced and expanded), and CAR T-cells expanded for 0, 10, or 17 days in culture. Black columns represent BTT-cells, and red columns represent FT6-treated T-cells. Data represent mean � S.E. of two independent experiments, triplicates. Ordinary one-way ANOVA (p �0.0001) with Bonferroni’s multiple comparison test comparing BT with FT6 categories for each T-cell population: **, p � 0.01; ***, p � 0.001.

Site-directed delivery of CAR T-cells

J. Biol. Chem. (2019) 294(48) 18465–18474 18469

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

Our data indicate that exofucosylation significantly im-proves marrow homing of CAR T-cells, as, at each cell dose, wefound higher numbers and higher proportions of FT6 CART-cells in marrow compared with injected BT cells (Fig. 5D, toppanels). Notably, the biggest difference in marrow infiltrationbetween FT6 and BT CAR T-cells was observed at cell dosesmuch lower than those routinely used in clinical applications(i.e. injection of 40,000 –100,000 cells/mouse (weight, 30 g) isreflective of doses of 1.08 –2.7 105/kg of body weight in ahuman (24)) (Fig. 5D, bottom panel). At these doses, weobserved �10-fold higher infiltration of exofucosylated cellscompared to unfucosylated cells. In other words, we found that

injection of 40,000 exofucosylated human CAR T-cells deliversthe same number of CAR T-cells into bone marrow as injectionof 400,000 buffer-treated cells. Importantly, the difference inmarrow infiltration between FT6 and BT cells is inversely pro-portional to cell dose, as, at higher doses (1–2 106 cells/mouse, i.e. 2.7–5.4 106/kg of body weight human dose), onlya 3.5 � 2.6 –fold higher frequency of exofucosylated cellswas observed in marrow compared to unfucosylated cells.Together, these findings support the notion that intravenousinfusion of high cell doses leads to saturation of tissue beds in astochastic fashion, i.e. regardless of molecular effectors of cellmigration, leading to nonspecific accumulation of adminis-

Figure 5. Exofucosylation significantly improves CAR T-cell homing to bone marrow. A, schematic of CAR T-cell homing experiment. 10- or 17-day CART-cells were either treated with buffer alone (BT) or exofucosylated using FT6 (FT6). These cells were injected through the retro-orbital plexus of NSG mice atvarying cell doses. 24 h after injection, mouse bone marrow (BM) was harvested and interrogated by flow cytometry. B, Number of cells in each category (BT orFT6) injected in individual mice, translated to number of cells per kilogram of body weight according to the following formula: human cell dose (cell numberper kilogram) � mouse cell dose (cell number per kilogram) (mouse Km/human Km) (24). C, representative dot plots showing flow cytometry analysis ofmouse bone marrow. Bone marrow cells were co-stained with antibodies against mouse CD45 (mCD45), human CD45 (hCD45), and human CD3 (hCD3).mCD45� and hCD45� cells (left) were subgated based on hCD3 expression (right) to identify hCD3 and hCD45 double-positive cells. Shown are bone marrowcells of mice receiving no human cell injection (top), 1 106 BT CAR T-cells (center), and 1 106 FT6-treated CAR T cells (bottom). Counting beads were usedto calculate the absolute number of human cells in the bone marrow. D, quantitation of CAR T-cell marrow homing, presented as a function of injected cellnumber. Total number of bone marrow–infiltrating T-cells (top left), percentage of human cells in mouse bone marrow (top right), and -fold difference in homedhuman cells in mouse bone marrow (bottom), calculated as follows: -fold difference � percentage of hCD3� hCD45� FT6 cells in mouse BM/percentage ofhCD3� hCD45� BT cells in mouse BM. Data are presented as mean � S.E. of three independent experiments. Each data point represents three to four mice.Ratio-paired t test comparing BT and FT6 conditions for each cell dose: *, p � 0.05; **, p � 0.01.

Site-directed delivery of CAR T-cells

18470 J. Biol. Chem. (2019) 294(48) 18465–18474

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

tered cells (25). In contrast, enforcing sLeX display by exofuco-sylation pilots CAR T-cell colonization preferentially intotissues whose endothelial beds express E-selectin. Thus, exofu-cosylation can potentially reduce CAR T-cell dosing by 10-fold.Notably, we did not observe any difference in spleen infiltrationbetween unfucosylated and exofucosylated CAR T-cells (Fig.S2), indicating that exofucosylation specifically increases CART-cell infiltration into E-selectin– bearing tissues. In particular,the results herein prompt future investigations using exofuco-sylated EGFR-directed CAR T-cells to assess their efficacy ineradication of osteotropic tumors that characteristically displayhigh levels of EGFR, such as in breast, lung, and prostate can-cers; notably, heightened EGFR expression is a driver of bonemetastasis of prostate cancer (26).

To determine whether exofusylation alters CAR T-cell cyto-toxicity, we investigated cytolysis by these cells in vitro (Fig. S3).We observed that CAR-T cells produce highly specific cytolysisof target U87 cells, which remains unaltered after exofucosyla-tion. Thus, the anti-tumor cytotoxicity of CAR T-cells is notmodified by cell surface glycoengineering.

In this study, we investigated the first step of CAR T-cellhoming, i.e. E-selectin–mediated recruitment of CAR T-cellsunder hemodynamic shear conditions, an important but hith-erto unexplored prerequisite for target-specific colonization ofCAR T-cells. Our study provides first evidence that CART-cells display an inability to engage E-selectin by virtue of thefact that these cells are deficient in cell surface sLeX display. Ourresults indicate that the absence of sLeX display on CAR T-cellsis not due to transduction of the CAR construct but a conse-quence of down-regulation of �(1,3)-FUTs during cell expan-sion, regardless of whether cells are cultured in FBS or HS.These results are in line with the findings of a prior study thatreported suppression of the �(1,3)-FUT enzyme FUT7 uponculturing human T-cells using conventional culture methods(27). To enforce E-selectin binding activity, we employed exo-fucosylation to custom-install sLeX on CAR T-cell surface, withresultant marked improvement in CAR T-cell homing to bonemarrow, an E-selectin– bearing target tissue. Notably, sLeX

installed via exofucosylation persists on the cell surface for �48h, after which the sLeX levels dissipate because of natural turn-over of cell surface glycoconjugates. Our central hypothesis isthat installation of sLeX on CAR T-cells via exofucosylation willprovide an initial boost in CAR T-cell targeting to E-selectin–bearing endothelial beds (as characteristic of native marrowendothelial beds and tumor endothelial beds) immediately afterinjection, and, this higher initial seeding of CAR T-cells withinthe tumor site will accentuate tumor elimination by CART-cells. Once CAR T-cells have infiltrated the tumor paren-chyma and come in contact with relevant lesional cells, theywill undergo intensive antigen-specific proliferation. Indeed,beyond induction of proliferation, the high affinity of the CARfor its cognate antigen will anchor the cells within the tumorparenchyma. However, alternate genetic approaches may alsobe utilized in order to achieve more stable sLeX display on thecell surface; e.g. using transfection of modified mRNA encodingFUT6 transcript (28), which allows for a non-permanent andnon-genome integrative gene expression lasting about 4 –5days post-transfection (19), one can achieve custom modifica-

tion of cell surface sLeX display for a longer duration. Further-more, co-expression of FT6 enzyme together with the CARconstruct, using a lentiviral delivery system, can create CART-cells with permanent display of cell surface sLeX.

Collectively, our findings reveal that cell surface glycoengi-neering can result in efficient and specific piloting of CART-cells into tissue sites that contain E-selectin– expressingendothelial beds. Importantly, apart from its constitutive ex-pression on marrow microvessels, E-selectin is also character-istically expressed by the microvascular endothelial cells of can-cers, and, thus, is strategically placed to recruit CAR T-cells.Notably, bone is a common metastastic site for a variety of solidmalignancies, and marrow microvessel expression of E-selectinis known to promote bone metastasis of cancer cells. Therefore,the ability of CAR T-cells to home to E-selectin– bearing sitessuch as marrow is critical for precise targeting of osteotropicmetastatic cancers such as prostate, breast, and lung adenocar-cinomas as well as for hematologic malignancies such as acuteleukemias and multiple myeloma. This improved homing maypermit more efficient localization of these immunotherapeuticcells at lesional tumor sites, achieving requisite tissue infiltra-tion with administration of fewer CAR T-cells. Thus, glycoen-gineering of the CAR T-cell surface could result in a signifi-cantly reduced burden of cell expansion, commensuratelyyielding less costly CAR T-cell therapeutics with higher clinicalefficacy.

Materials and methods

Study approval

Collection and use of human T-cells were in accordance withprocedures approved by institutional review board of the DanaFarber/Harvard Cancer Center (protocol 01-206) and Massa-chusetts General Hospital. The procedures included in thestudy abide by the Declaration of Helsinki principles. All animalstudies were in accordance with the National Institutes ofHealth guidelines for the care and use of animals underapproval of the institutional animal care and use committees ofBrigham and Women’s Hospital.

Cell culture

Human umbilical vein endothelial cells (HUVECs) were cul-tured in EBM2 medium (Lonza, Walkersville, MD) with 5%FBS.

Production and culture of huEGFR CAR T-cells

The huEGFRscFv-BBz chimeric antigen receptor wasdesigned based on the heavy and light chains of cetuximab toform a single-chain variable fragment (29, 30) that was fused toa portion of the extracellular and transmembrane domains ofhuman CD8�, followed by the intracellular domains of 4-1BBand CD3�. The bicistronic vector also encoded truncatedmCherry as a selectable marker and was placed following a T2Aribosomal skip sequence. The plasmid encoding huEGFRscFv-BBz-CAR was synthesized, and a third-generation lentiviruswas produced. T-cells were isolated from leukapheresis prod-ucts obtained from deidentified healthy donors and were stim-ulated in culture with Dynabeads Human T Activator CD3/

Site-directed delivery of CAR T-cells

J. Biol. Chem. (2019) 294(48) 18465–18474 18471

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

CD28 (Life Technologies) at a bead-to-cell ratio of 3:1. T-cellswere cultured in RPMI 1640 medium supplemented with either10% fetal bovine serum or with human AB serum, Hepes buffer(20 mM), penicillin (100 IU/ml), streptomycin (100 �g/ml), andIL-2 (20 IU/ml). One day following bead stimulation, T cellswere either transduced with CAR plasmid (CAR T) or leftuntransduced (UT) and then expanded for 10 days. After 10days of culture, a portion of the T-cells was debeaded and cryo-preserved; the remaining 10-day-expanded T cells were stimu-lated with irradiated U87 cells (a human glioma cell line thatexpresses EGFR), culture-expanded for an additional 7 days,and then cryopreserved until further use. Surface expression ofthe CAR was confirmed and quantified by flow cytometry basedon expression of mCherry; mCherry expression served as thereporter for identifying the extent of successful CAR transduc-tion in cells that had undergone transduction.

Antibodies

Antibodies used for flow cytometry were as follows:HECA452-FITC (catalog no. 321308, Biolegend), HECA452-Pacific Blue (catalog no. 321307, Biolegend) hCD45-FITC(clone HI130, catalog no. 304006, Biolegend), hCD3-PE-Cy7(clone OKT3, catalog no. 317334, Biolegend), mCD45-APC(30-F11, catalog no. 103112, Biolegend), hCD4 (clone OKT4,catalog no. 317416, Biolegend), mouse E-selectin–Ig chimera(catalog no. 575-ES, R&D Systems), anti-PSGL1 (clone KPL1,catalog no. 556053, BD Pharmingen), anti-CD43 (cloneL60, catalog no. 551457, BD Pharmingen), rat anti-mouse E-se-lectin (catalog no. MAB5751-500, R&D Systems), goat anti-ratIgG-HRP (catalog no. 3030-05, Southern Biotech), and goatanti-mouse-Ig-HRP (catalog no. 554002, BD Biosciences).

Flow cytometry and cell sorting

For flow cytometry analysis, cryopreserved T-cells were firstrevived and washed with PBS with 2% FBS. Cells were thenincubated with human Fc receptor blocking reagent TruStain(catalog no. 422302, Biolegend). 5 106 cells were then stainedwith 100 �l of appropriate antibody mixture prepared in PBSand 2% FBS. Staining was performed at 4 °C for 30 min. Stainedcells were then washed three times in PBS with 2% FBS andresuspended in 200 �l of PBS before analysis. Flow cytometrywas performed using BD FACS instruments (BD Biosciences).For multicolor experiments, compensation controls were setup using UltraComp eBeads (Invitrogen), and spectral overlapwas calculated using the automatic module of BD FACSDIVA.Flow cytometry data thus obtained were analyzed using FlowJosoftware (Tree Star, Ashland, OR). CAR T-cells and NT T-cellswere isolated using a BD flow-assisted cell sorter based onexpression of the mCherry reporter.

Exofucosylation

Native T-cells, and, 10-day-expanded or 17-day-expandedCAR T-cells were divided into two treatment groups. Onegroup was exofucosylated with FT6, and the other group wastreated with reaction buffer alone (BT). For exofucosylation,5 106 cells were incubated with 60 �l of reaction mixturecontaining 1 mM GDP-fucose and 60 �g/ml purified FT6enzyme (Warrior Therapeutics LLC, Sudbury, MA) in reaction

buffer at 37 °C for 1 h (19). For buffer treatment, cells wereincubated with reaction buffer alone (without enzyme or GDPfucose) at 37 °C for 1 h. Both groups were then washed with 1PBS prior to downstream analysis.

Preparation of cell lysates

Cell lysis was performed using a buffer containing 150 mM

NaCl, 50 mM Tris-HCI (pH 7.4), 0.02% NaN3, 20 mg/ml PMSF,2% Nonidet P-40, and protease inhibitor mixture (Roche). Aftersuspending the cell pellet in the aforementioned buffer, theresulting suspension was incubated on ice for 30 min andthen sonicated. The samples were then centrifuged at12,000 g for 5 min to remove particulate matter. Thesupernatant was then either boiled in reducing Laemmliloading buffer (Boston Bioproducts) to prepare for Westernblotting or used for immunoprecipitation.

Immunoprecipitation of E-selectin ligands

Cell lysates were precleared using protein G–agarose beads(Invitrogen) and then incubated with the murine E-selectin–human Fc chimera (E-Ig, Bio-Techne) in the presence of 2 mM

CaCl2. Immunoprecipitated glycoproteins were collected byboiling the beads in the presence of 2-mercaptoethanol inLaemmli loading buffer.

Western blotting

Reduced cell lysate protein samples or immunoprecipitationproducts were resolved on a 4%–20% SDS-PAGE gel (Bio-Rad)and then transferred onto a PVDF membrane (Bio-Rad). Themembrane was then blocked with 10% milk in TBS-T (TBS and0.1% Tween 20) buffer. For detection of E-selectin ligands,membranes were first incubated with E-Ig in TBS-T buffer con-taining 2 mM CaCl2 followed by incubation with rat anti-mouseE-selectin antibody (Bio-Techne) and then probed with HRP-conjugated goat anti-rat IgG antibody (Southern Biotech). Inother cases, membranes were incubated with monoclonal anti-bodies against either PSGL1 (clone KPL1, Biolegend), CD43(clone IG10, BD Biosciences), or CD44 (clone 156-3C11, CellSignaling), followed by incubation with HRP-conjugated goatanti-mouse IgG1 antibody (BD Biosciences). Protein bandswere detected by chemiluminescence using Lumi-Light sub-strate (Roche).

Quantitative RT-PCR

Total cellular RNA was reverse-transcribed using Super-Script VILO cDNA Conversion Kit (Invitrogen) according tothe following conditions: 25 °C for 10 min, 42 °C for 60 min, and85 °C for 5 min. Quantitative RT-PCR was performed with spe-cific primers to amplify target genes (listed in Table S1) usingSYBR Select Master Mix (Applied Biosystems, Foster City, CA)and the StepOne Plus PCR Detection System (Applied Biosys-tems) according to the following thermal cycling sequence: ini-tial activation at 95 °C for 20 s, followed by 40 cycles of 95 °C for10 s and 60 °C for 30 s. Expression of a particular gene in testcells (10-day or 17-day CAR T-cells) is presented as log2 -foldchange compared to expression of the said gene in nativehuman T-cells (untransduced and nonexpanded T-cells).

Site-directed delivery of CAR T-cells

18472 J. Biol. Chem. (2019) 294(48) 18465–18474

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

Microfluidic flow chamber experiment to evaluate tetheringand rolling function of CAR T-cells ex vivo

Microfluidic cell adhesion assay was performed using amicrofluidic flow chamber as described previously (31). For thisassay, HUVECs were cultured in 100-mm tissue culture dishesto obtain a uniform monolayer. HUVEC monolayers wereeither left unstimulated or stimulated to express E-selectin byincubating the cells with 40 ng/ml TNF� for 4 h at 37 °C. Themicrofluidic flow chamber was vacuum-sealed on the stimu-lated or unstimulated HUVEC monolayers and mounted over amicroscope stage. Native T-cells and culture-expanded NT ortransduced T-cells with or without exofucosylation were resus-pended at 2 106/ml cell concentration in HEPES buffer with2 mM CaCl2 and perfused into the microfluidic flow chamberusing a syringe pump at 2 dynes/cm2 wall shear stress. Cellularinteractions within the microfluidic chamber were recordedusing Moticam (Motic, Richmond, BC, Canada) camera and�Manager image acquisition software. Total rolling and adher-ent cells were enumerated using ImageJ analysis software.

Murine model to evaluate CAR T-cell homing

Age-matched NCG mice (NOD-Prkdcem26Cd52Il2rgem26Cd22/NjuCr) were purchased from Charles River (Wilmington MA).Buffer-treated or exofucosylated CAR T-cells were resus-pended in 1 PBS. The cells were then divided into seven dif-ferent cell doses as indicated in Fig. 3B, with each cell doseresuspended in a 100-�l injection volume. The cells were theninjected into the retro-orbital plexus of individual mice using a29-gauge hypodermic needle. 24 h after injection, mice weresacrificed, and bone marrow was harvested from two femoraand two tibiae of each mouse. The bone marrow cells weresubjected to red blood cell lysis, washed, and then stained withantibodies against mouse CD45, human CD45, and humanCD3. Bone marrow cells were then analyzed using flow cytom-etry to determine human cell infiltration within mouse bonemarrow.

Cytotoxicity assay

Human EGFR CAR T-cells were co-cultured with click bee-tle green luciferase– expressing U87 cells as targets at the indi-cated ratios (Fig. S3) for 16 h. Luciferase activity was measuredwith a Synergy Neo2 luminescence microplate reader (Biotek),and the percentage of specific lysis was calculated utilizing thefollowing equation: percent specific lysis � (total relative lightunits/target cell specific relative light units) 100.

Statistics

Data are reported as mean � S.E. The p values were calcu-lated using a ratio paired t test or ordinary one-way ANOVAwith Sidak’s multiple comparison test, as indicated in the figurelegends. p � 0.05 is considered significant.

Author contributions—R. S. conceptualization; N. M., M. S., andA. P. C. data curation; N. M. and M. S. formal analysis; N. M. andR. S. investigation; N. M. visualization; N. M., A. P. C., and M. V. M.methodology; N. M. and R.S. writing-original draft; N. M., A. P. C.,and R. S. writing-review and editing; M. V. M. and R. S. resources;R. S. funding acquisition; R. S. validation.

Acknowledgment—We thank Kyle C. Martin for assistance withmouse studies.

References1. Hay, K. A., Hanafi, L. A., Li, D., Gust, J., Liles, W. C., Wurfel, M. M., López,

J. A., Chen, J., Chung, D., Harju-Baker, S., Cherian, S., Chen X., Riddell,S. R., Maloney, D. G., and Turtle C. J. (2017) Kinetics and biomarkers ofsevere cytokine release syndrome after CD19 chimeric antigen receptor-modified T-cell therapy. Blood 130, 2295–2306 CrossRef Medline

2. Hill, J. A., Li, D., Hay, K. A., Green, M. L., Cherian, S., Chen, X., Riddell,S. R., Maloney, D. G., Boeckh, M., and Turtle, C. J. (2018) Infectious com-plications of CD19-targeted chimeric antigen receptor-modified T-cellimmunotherapy. Blood 131, 121–130 CrossRef Medline

3. Morgan, R. A., Yang, J. C., Kitano, M., Dudley, M. E., Laurencot, C. M., andRosenberg, S. A. (2010) Case report of a serious adverse event followingthe administration of T cells transduced with a chimeric antigen receptorrecognizing erbb2. Mol. Ther. 18, 843– 851 CrossRef Medline

4. Mukai, S., Kjaergaard, J., Shu, S., and Plautz, G. E. (1999) Infiltration oftumors by systemically transferred tumor-reactive T lymphocytes is re-quired for antitumor efficacy. Cancer Res. 59, 5245–5249 Medline

5. Sackstein, R., Schatton, T., and Barthel, S. R. (2017) T-lymphocyte hom-ing: an underappreciated yet critical hurdle for successful cancer immu-notherapy. Lab. Invest. 97, 669 – 697 CrossRef Medline

6. Sackstein, R. (2009) Glycosyltransferase-programmed stereosubstitution(GPS) to create HCELL: engineering a roadmap for cell migration. Immu-nol. Rev. 230, 51–74 CrossRef Medline

7. Schweitzer, K. M., Dräger, A. M., van der Valk, P., Thijsen, S. F., Zeven-bergen, A., Theijsmeijer, A. P., van der Schoot, C. E., and Langenhuijsen,M. M. (1996) Constitutive expression of E-selectin and vascular cell adhe-sion molecule-1 on endothelial cells of hematopoietic tissues. Am. J.Pathol. 148, 165–175 Medline

8. Tamaru, M., Tomura, K., Sakamoto, S., Tezuka, K., Tamatani, T., andNarumi, S. (1998) Interleukin-1� induces tissue- and cell type-specificexpression of adhesion molecules in vivo. Arterioscler. Thromb. Vasc. Biol.18, 1292–1303 CrossRef Medline

9. Kannagi, R., Izawa, M., Koike, T., Miyazaki, K., and Kimura, N. (2004)Carbohydrate-mediated cell adhesion in cancer metastasis and angiogen-esis. Cancer Science 95, 377–384 CrossRef Medline

10. Esposito, M., Mondal, N., Greco, T. M., Wei, Y., Spadazzi, C., Lin, S. C.,Zheng, H., Cheung, C., Magnani, J. L., Lin, S. H., Cristea, I. M., Sackstein,R., and Kang, Y. (2019) Bone vascular niche e-selectin induces mesenchy-mal-epithelial transition and WNT activation in cancer cells to promotebone metastasis. Nat. Cell Biol. 21, 627– 639 CrossRef Medline

11. Silva, M., Fung, R. K. F., Donnelly, C. B., Videira, P. A., and Sackstein, R.(2017) Cell-specific variation in E-selectin ligand expression among hu-man peripheral blood mononuclear cells: implications for immunosur-veillance and pathobiology. J. Immunol. 198, 3576 –3587 CrossRefMedline

12. Mondal, N., Stolfa, G., Antonopoulos, A., Zhu, Y., Wang, S. S., Buffone, A.,Jr., Atilla-Gokcumen, G. E., Haslam, S. M., Dell, A., and Neelamegham, S.(2016) Glycosphingolipids on human myeloid cells stabilize E-selectin-de-pendent rolling in the multistep leukocyte adhesion cascade. Arterioscler.Thromb. Vasc. Biol. 36, 718 –727 CrossRef Medline

13. Seshacharyulu, P., Ponnusamy, M. P., Haridas, D., Jain, M., Ganti, A. K.,and Batra, S. K. (2012) Targeting the EGFR signaling pathway in cancertherapy. Expert Opin. Ther. Targets 16, 15–31 CrossRef Medline

14. Mellinghoff, I. K., Wang, M. Y., Vivanco, I., Haas-Kogan, D. A., Zhu, S.,Dia, E. Q., Lu, K. V., Yoshimoto, K., Huang, J. H., Chute, D. J., Riggs, B. L.,Horvath, S., Liau, L. M., Cavenee, W. K., Rao, P. N., et al. (2005) Moleculardeterminants of the response of glioblastomas to EGFR kinase inhibitors.N. Engl. J. Med. 353, 2012–2024 CrossRef Medline

15. Johnson, L. A., Scholler, J., Ohkuri, T., Kosaka, A., Patel, P. R., McGettigan,S. E., Nace, A. K., Dentchev, T., Thekkat, P., Loew, A., Boesteanu, A. C.,Cogdill, A. P., Chen, T., Fraietta, J. A., Kloss, C. C., et al. (2015) Rationaldevelopment and characterization of humanized anti-EGFR variant III

Site-directed delivery of CAR T-cells

J. Biol. Chem. (2019) 294(48) 18465–18474 18473

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

chimeric antigen receptor T cells for glioblastoma. Sci. Transl. Med. 7,275ra22 CrossRef Medline

16. Sackstein, R., Merzaban, J. S., Cain, D. W., Dagia, N. M., Spencer, J. A., Lin,C. P., and Wohlgemuth, R. (2008) Ex vivo glycan engineering of CD44programs human multipotent mesenchymal stromal cell trafficking tobone. Nat. Med. 14, 181–187 CrossRef Medline

17. Ballermann, B. J., Dardik, A., Eng, E., and Liu, A. (1998) Shear stress andthe endothelium. Kidney Int. Suppl. 67, S100-S108 Medline

18. Dagia, N. M., Gadhoum, S. Z., Knoblauch, C. A., Spencer, J. A., Zamiri, P.,Lin, C. P., and Sackstein, R. (2006) G-CSF induces E-selectin ligand ex-pression on human myeloid cells. Nat. Med. 12, 1185–1190 CrossRefMedline

19. Dykstra, B., Lee, J., Mortensen, L. J., Yu, H., Wu, Z. L., Lin, C. P., Rossi, D. J.,and Sackstein, R. (2016) Glycoengineering of E-selectin ligands by intra-cellular versus extracellular fucosylation differentially affects osteotro-pism of human mesenchymal stem cells. Stem Cells 34, 2501–2511CrossRef Medline

20. Thankamony, S. P., and Sackstein, R. (2011) Enforced hematopoietic cellE- and L-selectin ligand (HCELL) expression primes transendothelial mi-gration of human mesenchymal stem cells. Proc. Natl. Acad. Sci. U.S.A.108, 2258 –2263 CrossRef Medline

21. Abdi, R., Moore, R., Sakai, S., Donnelly, C. B., Mounayar, M., and Sack-stein, R. (2015) HCELL expression on murine MSC licenses pancreatot-ropism and confers durable reversal of autoimmune diabetes in NODmice. Stem Cells 33, 1523–1531 Medline

22. Lee, J., Dykstra, B., Spencer, J. A., Kenney, L. L., Greiner, D. L., Shultz, L. D.,Brehm, M. A., Lin, C. P., Sackstein, R., and Rossi, D. J. (2017)mRNA-mediated glycoengineering ameliorates deficient homing of hu-man stem cell-derived hematopoietic progenitors. J. Clin. Invest. 127,2433–2437 CrossRef Medline

23. Videira, P. A., Silva, M., Martin, K. C., and Sackstein, R. (2018) Ligation ofthe CD44 glycoform HCELL on culture-expanded human monocyte-de-rived dendritic cells programs transendothelial migration. J. Immunol.201, 1030 –1043 CrossRef Medline

24. Nair, A. B., and Jacob, S. (2016) A simple practice guide for dose conver-sion between animals and human. J. Basic Clin. Pharm. 7, 27–31 CrossRefMedline

25. Sackstein, R. (2018) The first step in adoptive cell immunotherapeutics:assuring cell delivery via glycoengineering. Front. Immunol. 9, 3084Medline

26. Day, K. C., Lorenzatti Hiles, G., Kozminsky, M., Dawsey, S. J., Paul, A.,Broses, L. J., Shah, R., Kunja, L. P., Hall, C., Palanisamy, N., Daignault-Newton, S., El-Sawy, L., Wilson, S. J., Chou, A., Ignatoski, K. W., et al.(2017) HER2 and EGFR overexpression support metastatic progression ofprostate cancer to bone. Cancer Res. 77, 74 – 85 CrossRef Medline

27. Knibbs, R. N., Craig, R. A., Natsuka, S., Chang, A., Cameron, M., Lowe,J. B., and Stoolman, L. M. (1996) The fucosyltransferase FucT-VII regu-lates E-selectin ligand synthesis in human T cells. J. Cell Biol. 133,911–920 CrossRef Medline

28. Mondal, N., Dykstra, B., Lee, J., Ashline, D. J., Reinhold, V. N., Rossi, D. J.,and Sackstein, R. (2018) Distinct human �(1,3)-fucosyltransferases driveLewis-X/sialyl Lewis-X assembly in human cells. J. Biol. Chem. 293,7300 –7314 CrossRef Medline

29. Caruso, H. G., Torikai, H., Zhang, L., Maiti, S., Dai, J., Do, K. A., Singh, H.,Huls, H., Lee, D. A., Champlin, R. E., Heimberger, A. B., and Cooper, L. J.(2016) Redirecting T-cell specificity to EGFR using mRNA to self-limitexpression of chimeric antigen receptor. J. Immunother. 39, 205–217CrossRef Medline

30. Banisadr, A., Safdari, Y., Kianmehr, A., and Pourafshar, M. (2018) Produc-tion of a germline-humanized cetuximab scFv and evaluation of its activityin recognizing EGFR-overexpressing cancer cells. Hum. Vaccin. Immu-nother. 14, 856 – 863 CrossRef Medline

31. Mondal, N., Buffone, A., Jr., Stolfa, G., Antonopoulos, A., Lau, J. T.,Haslam, S. M., Dell, A., and Neelamegham, S. (2015) St3gal-4 is theprimary sialyltransferase regulating the synthesis of E-, P-, and L-se-lectin ligands on human myeloid leukocytes. Blood 125, 687– 696CrossRef Medline

Site-directed delivery of CAR T-cells

18474 J. Biol. Chem. (2019) 294(48) 18465–18474

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

SacksteinNandini Mondal, Mariana Silva, Ana P. Castano, Marcela V. Maus and Robert

bindingGlycoengineering of chimeric antigen receptor (CAR) T-cells to enforce E-selectin

doi: 10.1074/jbc.RA119.011134 originally published online October 18, 20192019, 294:18465-18474.J. Biol. Chem. 

  10.1074/jbc.RA119.011134Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/294/48/18465.full.html#ref-list-1

This article cites 31 references, 13 of which can be accessed free at

at Biom

edical Library, U

CSD

on January 20, 2020http://w

ww

.jbc.org/D

ownloaded from

Supporting Information for

Glycoengineering of Chimeric Antigen Receptor (CAR) T-cells to Enforce E-selectin Binding

Nandini Mondal, Mariana Silva, Ana P. Castano, Marcela V. Maus, and Robert Sackstein

Corresponding Author: Robert SacksteinE-mail: sackstein@fiu.edu

This PDF file includes:

Supplementary Figures. S1 to S3Supplementary Table S1

Table S1. List of primers for quantitative RT-PCR analysis

Enzymes Forward primer (5’-3’) Reverse Primer (5’-3’) Source

1. FUT3 GCCGACCGCAAGGTGTAC TGACTTAGGGTTGGACATGATATCC Higai, Ishihara et al. 20062. FUT5 ACCTGAGCTACTTTCACTGGCG TCAGGTGAACCAAGCCGCTATG Origene3. FUT6 CCGACTACATCACCGAGAAGCT GAACCTCTCGTAGTTGCTTCTGC Origene4. FUT7 GAATGAGAGCCGATACCAACGC TAGCGGTCACAGATGGCACAGA Origene

2 of 5 Nandini Mondal, Mariana Silva, Ana P. Castano, Marcela V. Maus, and Robert Sackstein

Cell c

ou

nt

sLeX (HECA452)

Cell c

ou

nt

Cell c

ou

nt

Cell c

ou

nt

sLeX (HECA452)

sLeX (HECA452) sLeX (HECA452)

CD8+ CD4+

mCherry- mCherry-

mCherry- mCherry+ mCherry+mCherry-

mCherry- mCherry+ mCherry+mCherry-

BT

FT6

i ii iii iv

i ii iii iv

250150

100

75

50

37

NT

T-cells

CAR

T-cells

CAR

T-cells

17-Day10-Day

GAPDH

A D

B

E

CE-selectin-Ig

Immunoprecipitation

250

150

10075

250

150

100

75

Mo

l. W

t

(KD

a)

Mo

l. W

t

(KD

a)

PSGL1(CLA) CD43E/HCELL

CD43E

CD44 (HCELL)

N T C A R T N T C A R T

-1 0

-5

0

5

mR

NA

le

ve

ls

(Lo

g2

-fo

ld c

ha

ng

e)

1 0 -D a y 1 7 -D a y

N T C A R T N T C A R T

-1 0

-5

0

5

1 0 -D a y 1 7 -D a y

F G

Cell c

ou

nt

sLeX (HECA452)

Cell c

ou

nt

Cell c

ou

nt

Cell c

ou

nt

sLeX (HECA452)

sLeX (HECA452) sLeX (HECA452)

CD8+ CD4+

mCherry- mCherry-

mCherry- mCherry+ mCherry+mCherry-

mCherry- mCherry+ mCherry+mCherry-

BT

FT6

i ii iii iv

i ii iii iv

250150

100

75

50

37

NT

T-cells

CAR

T-cells

CAR

T-cells

17-Day10-Day

GAPDH

A D

B

E

CE-selectin-Ig

Immunoprecipitation

250

150

10075

250

150

100

75

Mo

l. W

t

(KD

a)

Mo

l. W

t

(KD

a)

PSGL1(CLA) CD43E/HCELL

CD43E

CD44 (HCELL)

N T C A R T N T C A R T

-1 0

-5

0

5m

RN

A l

ev

els

(Lo

g2

-fo

ld c

ha

ng

e)

1 0 -D a y 1 7 -D a y

N T C A R T N T C A R T

-1 0

-5

0

5

1 0 -D a y 1 7 -D a y

F G

mR

NA

lev

els

(L

og

2-F

old

ch

an

ge)

Cell c

ou

nt

sLeX (HECA452)

Cell c

ou

nt

Cell c

ou

nt

Cell c

ou

nt

sLeX (HECA452)

sLeX (HECA452) sLeX (HECA452)

CD8+ CD4+

mCherry- mCherry-

mCherry- mCherry+ mCherry+mCherry-

mCherry- mCherry+ mCherry+mCherry-

BT

FT6

i ii iii iv

i ii iii iv

250150

100

75

50

37

NT

T-cells

CAR

T-cells

CAR

T-cells

17-Day10-Day

GAPDH

A D

B

E

CE-selectin-Ig

Immunoprecipitation

250

150

10075

250

150

100

75

Mo

l. W

t

(KD

a)

Mo

l. W

t

(KD

a)

PSGL1(CLA) CD43E/HCELL

CD43E

CD44 (HCELL)

N T C A R T N T C A R T

-1 0

-5

0

5

mR

NA

le

ve

ls

(Lo

g2

-fo

ld c

ha

ng

e)

1 0 -D a y 1 7 -D a y

N T C A R T N T C A R T

-1 0

-5

0

5

1 0 -D a y 1 7 -D a y

F GD

E F

Cell c

ou

nt

sLeX (HECA452)

Cell c

ou

nt

Cell c

ou

nt

Cell c

ou

nt

sLeX (HECA452)

sLeX (HECA452) sLeX (HECA452)

CD8+ CD4+

mCherry- mCherry-

mCherry- mCherry+ mCherry+mCherry-

mCherry- mCherry+ mCherry+mCherry-

BT

FT6

i ii iii iv

i ii iii iv

250150

100

75

50

37

NT

T-cells

CAR

T-cells

CAR

T-cells

17-Day10-Day

GAPDH

A D

B

E

CE-selectin-Ig

Immunoprecipitation

250

150

10075

250

150

100

75

Mo

l. W

t

(KD

a)

Mo

l. W

t

(KD

a)

PSGL1(CLA) CD43E/HCELL

CD43E

CD44 (HCELL)

N T C A R T N T C A R T

-1 0

-5

0

5

mR

NA

le

ve

ls

(Lo

g2

-fo

ld c

ha

ng

e)

1 0 -D a y 1 7 -D a y

N T C A R T N T C A R T

-1 0

-5

0

5

1 0 -D a y 1 7 -D a y

F G

Cell c

ou

nt

sLeX (HECA452)

Cell c

ou

nt

Cell c

ou

nt

Cell c

ou

nt

sLeX (HECA452)

sLeX (HECA452) sLeX (HECA452)

CD8+ CD4+

mCherry- mCherry-

mCherry- mCherry+ mCherry+mCherry-

mCherry- mCherry+ mCherry+mCherry-

BT

FT6

i ii iii iv

i ii iii iv

250150

100

75

50

37

NT

T-cells

CAR

T-cells

CAR

T-cells

17-Day10-Day

GAPDH

A D

B

E

CE-selectin-Ig

Immunoprecipitation

250

150

10075

250

150

100

75

Mo

l. W

t

(KD

a)

Mo

l. W

t

(KD

a)

PSGL1(CLA) CD43E/HCELL

CD43E

CD44 (HCELL)

N T C A R T N T C A R T

-1 0

-5

0

5

mR

NA

le

ve

ls

(Lo

g2

-fo

ld c

ha

ng

e)

1 0 -D a y 1 7 -D a y

N T C A R T N T C A R T

-1 0

-5

0

5

1 0 -D a y 1 7 -D a y

F G

D

Fig. S1. Culture expansion augments cell surface type 2 sialylLacNAc acceptors on T-cell surface. (A) Representative histograms of sLeX expression on untransducedCD8+ (left) and CD4+ (Right) T-cells. (B) Representative histograms of sLeX expression on 10-day expanded (i) CD8+ non-transduced T-cells, (ii) CD8+ CAR T-cells, (iii)CD4+ non-transduced T-cells, and (iv) CD4+ CAR T-cells. (C) Representative histogram of sLeX expression on 17-day expanded (i) CD8+ non-transduced T-cells, (ii) CD8+

CAR T-cells, (iii) CD4+ non-transduced T-cells, and (iv) CD4+ CAR T-cells. Black histograms represent buffer treated and red histograms present FT6-exofucosylated cells. (D)Western blot analysis using E-selectin-Ig chimera as a probe, of buffer treated (BT) or FT6 exofucosylated (FT6) 10-day expanded non-transduced (NT) or CAR T-cells and17-day expanded CAR T-cells. GAPDH staining was used as loading control. Numbers on the left indicate molecular weight in KDa.

Nandini Mondal, Mariana Silva, Ana P. Castano, Marcela V. Maus, and Robert Sackstein 3 of 5

No Injection BT FT6

0 .0 1 0 .1 1 1 00 .0 0 0 1

0 .0 0 1

0 .0 1

0 .1

1

C e ll N u m b e r in je c te d

p e r m o u s e (M il l io n )

hC

D3

+ c

ell

s i

n s

ple

en

(% m

CD

45

+ c

ell

s)

A B

mC

D4

5

hCD3

0 .0 1 0 .1 1 1 00 .0 0 0 1

0 .0 0 1

0 .0 1

0 .1

1

C e ll N u m b e r in je c te d

p e r m o u s e (M illio n )

hC

D3

+ c

ell

s i

n s

ple

en

(% m

CD

45

+ c

ell

s)

F T 6

B T

Fig. S2. CAR T-cell infiltration within spleen of systemically-administered cells is not altered upon exofucosylation. (A) Representative flow cytometry dot plotspresenting mouse CD45 and human CD3 staining of spleen cells harvested from NSG mice receiving either no human cell injection (left panel), buffer treated human CART-cell injection (BT, middle panel), or FT6-treated human CAR T-cell injection (FT6, right panel). (B) Infiltration of human CD3+ cells in mouse spleen, quantified as % of mouseCD45+ cells.

4 of 5 Nandini Mondal, Mariana Silva, Ana P. Castano, Marcela V. Maus, and Robert Sackstein

9:1

3:1

1:1

1:3

1:9

1:2

7

1:8

1

0

2 0

4 0

6 0

8 0

1 0 0

E ffe c to r :T a rg e t

% S

pe

cif

ic L

ys

is

U N U N

1 0 d a y C A R 1 0 d a y C A R

1 7 d a y C A R 1 7 d a y C A R

B T F T 6

UT UT

Fig. S3. Exofucosylation does not alter cytotoxicity of CAR T-cells. Untransduced T-cells (UT) and CAR T-cells were either treated with buffer alone, or incubated withFT6 and GDP-fucose for 1 hour at 37oC. These cells were co-cultured for 16 hours with target cells (beetle green (CBG) luciferase-expressing U87 cell line) at multiple effector:target ratios. Cell lysis was assessed by measuring Luciferase activity. Line plot presents percent specific cell lysis calculated as follows. % specific lysis = (total RLU / targetcells only RLU) x100. RLU= Relative luminometer units.

Nandini Mondal, Mariana Silva, Ana P. Castano, Marcela V. Maus, and Robert Sackstein 5 of 5