Autologous, lentivirus-modified, T-rapa cell "micropharmacies" for lysosomal storage disorders

Abstract

T cells are the current choice for many cell therapy applications. They are relatively easy to access, expand in culture, and genetically modify. Rapamycin-conditioning ex vivo reprograms T cells, increasing their memory properties and capacity for survival, while reducing inflammatory potential and the amount of preparative conditioning required for engraftment. Rapamycin-conditioned T cells have been tested in patients and deemed to be safe to administer in numerous settings, with reduced occurrence of infusion-related adverse events. We demonstrate that ex vivo lentivirus-modified, rapamycin-conditioned CD4⁺ T cells can also act as next-generation cellular delivery vehicles-that is, "micropharmacies"-to disseminate corrective enzymes for multiple lysosomal storage disorders. We evaluated the therapeutic potential of this treatment platform for Fabry, Gaucher, Farber, and Pompe diseases in vitro and in vivo. For example, such micropharmacies expressing α-galactosidase A for treatment of Fabry disease were transplanted in mice where they provided functional enzyme in key affected tissues such as kidney and heart, facilitating clearance of pathogenic substrate after a single administration.

Keywords: T cells; gene therapy; lentivirus; lysosomal storage disorders.

Figure EV1. Supranormal intracellular and secreted lysosomal hydrolase enzyme activity from lentivirus‐transduced HEK293T and Jurkat cells

1. A

   Schematic of the modified HIV‐1‐derived lentiviral backbone used in this study.

2. B, C

   Lentiviral vectors (LVs) were packaged and used to transduce HEK293T cells—vector copy number per genome (VCN/g) for each HEK293T line are indicated in (B). Enzyme activities in lysates (intracellular; B) and supernatants (secreted; C) were measured.

3. D, E

   LVs were used to transduce Jurkat cells—VCN/g for each Jurkat line are indicated in (D). Enzyme activities in lysates (intracellular; D) and supernatant (secreted; E) enzyme activities were measured.

Data information: Activities are reported as a mean of n = 3 seeded wells, error bars are standard deviation, and the y‐axis is logarithmic to base 10 to highlight activities in controls. Two‐tailed Student's t‐tests were used to compare activities in NT versus LV‐transduced cells, ***P < 0.001. Abbreviations—LTR: long‐terminal repeat; SIN/LTR: self‐inactivating LTR; SD: splice donor; SA: splice acceptor; Ψ: Retroviral Psi packaging element; ΔGAG: modified lentiviral Gag response element; RRE: Rev response element; cPPT: central polypurine tract; EF1α: elongation factor 1 α (promoter); WPRE: woodchuck hepatitis post‐translational regulatory element; GLA: α‐galactosidase A (α‐gal A); GBA: β‐glucocerebrosidase (GCase); ASAH1: acid ceramidase (ACDase); GAA: acid α‐glucosidase; NT: non‐transduced; LV: lentiviral transduced.

Figure 1. Transduced healthy (HDo) and Fabry donor (FDo)‐derived T‐Rapa “micropharmacies” (TRaMs) secrete functional lysosomal enzymes

1. A

   Schematic of TRaM production. CD4⁺ T cells were enriched by positive selection from leukaphereses of multiple HDos and 3 FDos—details are provided in Materials and Methods.

2. B

   Enzyme compartments that are measured in C‐I and Fig EV4G; intracellular, total secreted and colloidal particles (enclosed in red and yellow circles).

3. C

   HDo‐TRaMs were engineered with vectors relevant to treatment of Fabry, Gaucher, Farber, and Pompe diseases. Intracellular lysosomal enzyme specific activities were measured in transduced T‐Rapa and controls. Transductions varied across vectors/donors, as indicated.

4. D, E

   Enzyme activity was assessed in transduced T‐Rapa supernatants compared to controls in activated/dividing (D) and quiescent/resting (E) states.

5. F

   Cells from all FDo were stably transduced, as indicated. Intracellular α‐galactosidase A (α‐gal A)‐specific activities were determined.

6. G, H

   Secreted α‐gal A activity was measured from FDo‐derived TRaMs in activated (G), and quiescent/resting (H) states.

7. I

   α‐Gal A‐specific activity in Fabry patient‐derived skin fibroblasts after 6 h of exposure to conditioned media from TRaMs in the presence or absence of 1 mM M6P.

Data information: Activities in (C–H) are reported as a mean of n = 3 seeded wells per donor, error bars are standard deviation. Data for (C–E) are representative of at least two donors. Activities were compared between matched transduced and control cells using pairwise two‐tailed Student's t‐tests, ***P < 0.001 and **P < 0.01. Activities in (I) are reported as a mean of n = 3 independent treatments per condition, error bars are standard deviation. Two‐way analysis of variance with Tukey's multiple comparison test was used to compare activities, **P < 0.01 and *P < 0.05. Abbreviations—VCN/g: vector copy number per genome; GLA: α‐gal A; GBA: β‐glucocerebrosidase (GCase); ASAH1: acid ceramidase (ACDase); GAA: acid α‐glucosidase; NT: non‐transduced; LV: lentiviral transduced; CTRL: vehicle control.

Figure EV2. Rapamycin‐conditioned T cells are resistant to cryogenic stress, regardless of transduction

1. Viability of rapamycin‐conditioned T cells was compared to control T cells after a freeze–thaw cycle by trypan blue exclusion.

2. Rapamycin‐conditioned or control cells were thawed and seeded for culture with CD3/CD28 T‐Activator beads. Live cells, as determined by trypan blue exclusion, were counted every day.

Data information: Viability in (A) is shown as a mean of n = 3 donor cells, error bars are standard error of the mean. Two‐tailed Student's t‐test was used to compare viability between rapamycin‐conditioned (T‐Rapa) and control cells, ***P < 0.001 and **P < 0.01. For comparison, data in (B) is presented as mean of fold‐change from original seeding number of n = 2 donors each with n = 2 replicates, error bars are the range obtained from samples from two donors. Two‐way analysis of variance with Bonferroni's multiple comparison test showed significantly higher fold‐change of live cells at days 5 and 6, *P < 0.05. Abbreviations—NT: non‐transduced; LV: lentivirus‐modified.

Figure EV3. Rapamycin treatment has comparable low‐depth inhibition of the mTOR pathway in Healthy (HDo) and Fabry donor (FDo) T cells, and overexpression of α‐gal A does not suppress T cell function

1. CD4⁺ T cells from HDo or FDo were cultured with CD3/CD28 T‐Activator beads and 1 µM rapamycin for 3 days and harvested to perform Western blots for markers of mTOR inhibition.

2. Relative expression of phosphorylated mTOR targets were determined from densitometry of Western blots; intensities were normalized to β‐Actin, then to total protein, then to results from control cells.

3. FDo‐ and HDo‐derived TRaMs were seeded with CD3/CD28 T‐Activator beads for 24 h and secreted cytokines were measured by a Luminex assay using a custom multiplex panel.

4. Transduced FDo TRaMs and controls (from three independent donors) were seeded with CD3/CD28 T‐Activator beads for 2 days. Live cells were counted by trypan blue exclusion.

Data information: Bars in (B) represent means of relative expression from n = 2 individual replicates of cells from single donors. Error bars have been omitted; values vary by 0.4–25% for the control‐treated groups, and 1–10% for the rapamycin‐treated groups. Two‐way analysis of variance with Fisher's LSD test was used to compare expression, **P < 0.01 and *P < 0.05. Fold‐changes in (C) were calculated by normalizing cytokine concentrations to the average of matched non‐transduced (NT) controls. Mean fold‐change is plotted for n = 3 technical replicates of 3 sets of FDo and 1 set of HDo cells, error bars are standard error of the mean. A two‐way analysis of variance with Tukey's multiple comparison test was used to compare cytokine levels between groups; no significant changes were found. Data in (D) is shown as mean fold‐change from day 0, error bars are standard error of the mean. A two‐tailed Student's t‐test was used to compare growth of transduced and control cells in response to short‐term stimulation. Abbreviations—TRaMs: T‐Rapa micropharmacies; LV/GLA: lentivirus transduced to express α‐galactosidase A; IL: interleukin; IFN: interferon; GM‐CSF: granulocyte‐monocyte colony‐stimulating factor; TNF: tumor necrosis factor.

Figure EV4. LV/GLA secreted enzyme is mature, predominantly soluble, and can be taken up by Fabry fibroblasts

1. A

   Supernatants from culture of NT and LV/GLA HDo‐TRaMs, and recombinant α‐gal A were digested with the indicated glycosidases and Western blot with anti‐α‐gal A antibody was performed.

2. B

   α‐Gal A specific activity in Fabry patient‐derived skin fibroblasts after 6 h of exposure to conditioned media from HDo‐TRaMs in the presence or absence of 1 and 10 mM M6P.

3. C

   Schematic of the series of centrifugation steps used to clear FDo‐TRaM conditioned media of particulates.

4. D

   Enzyme activity was measured in conditioned media before (input) and after (particle‐depleted) ultracentrifugation.

5. E, F

   Nanoparticle tracking analysis was used to measure size distribution (E) and concentration (F) of colloidal particles collected after the final ultracentrifugation.

6. G

   α‐Gal A specific activities in colloidal particles isolated from conditioned media by ultracentrifugation were determined.

7. H

   Immunofluorescent microscopy of Fabry patient‐derived fibroblasts after culture with NT or LV/GLA 293T conditioned supernatant. For each condition, a representative image is shown with enlarged section from white box to the right. White arrows indicate cells in which enzyme has been taken up and is co‐localized with lysosomal staining.

Data information: Activities in (B) are reported as a mean of n = 3 independent treatments per condition, error bars are standard deviation. Activities in (B) were compared using a two‐way analysis of variance with Tukey's multiple comparison test. Activities in (D) are reported as a mean of n = 3 depletions per donor, error bars are standard deviation. Particle‐depleted activities in (D) were normalized to input activities and compared using a one‐way analysis of variance with Tukey's multiple comparison test. Concentrations of particles of each size were converted to a percentage of total particles for comparison between donors in (E); data are plotted as a distribution of means of n = 3 replicates of 3 donors for each of NT and LV/GLA, shown with a dark gray or green line. Standard error of the mean is shown as shading extending around these lines. Distributions were compared using a two‐way analysis of variance. Data in (F) is plotted as means of n = 3 counts for each donor, with standard error of the mean. NT and LV/GLA concentrations were compared using a one‐way analysis of variance with Tukey's multiple comparison test. Activities in (G) are reported as a mean of n = 3 isolations for each donor, error bars are standard deviation. Activities were compared between particles isolated from matched transduced and control cells using pairwise two‐tailed Student's t‐tests, ***P < 0.001. Scale bars in (H) represent 10 µm. Abbreviations—TRaMs: T‐Rapa micropharmacies; HDo and FDo: Healthy and Fabry patient‐derived cells, respectively; GLA: α‐galactosidase A (α‐gal A); NT: non‐transduced; LV/GLA: lentivirus transduced to express α‐galactosidase A; CTRL: vehicle control; PNGase F: Peptide‐N‐Glycosidase F; Endo H: Endoglycosidase H; DAPI: 4',6‐diamidino‐2‐phenylindole; LAMP1: Lysosomal‐associated membrane protein 1.

Figure 2. TRaMs result in elevated enzyme expression following xenotransplantation

1. A

   Schematic of the transplant protocol optimized for T‐Rapa cells. NOD/SCID/Fabry (NSF) mice were conditioned and TRaMs were administered intravenously—details are indicated in Materials and Methods. Mice were euthanized 4 weeks post‐infusion.

2. B

   Engraftment was evaluated by flow cytometry for human CD3 expression.

3. C, D

   α‐Gal A activity in plasma, and specific activities in liver, spleen, heart, and kidneys were measured after transplant of transduced healthy donor (HDo)‐derived (C) or Fabry donor (FDo)‐derived (D) TRaMs, non‐transduced (NT) cells, or in sham‐treated NOD/SCID (NS) and NSF mice.

Data information: %CD3 of lymphocyte‐gated parent population in (B) and enzyme activities in (C, D) are reported as values for individual mice (n = 4) with a line indicating the mean, and error bars showing standard error of the mean. The y‐axis is logarithmic to base 10 in (C, D) to highlight activities in controls. One‐way analyses of variance with Tukey's multiple comparison tests were used to compare activities between groups, ***P < 0.001, **P < 0.01, *P < 0.05, and ns P > 0.05. Abbreviations—TRaMs: T‐Rapa micropharmacies; LV/GLA: lentivirus transduced to express α‐gal A; VCN/g: vector copy number per genome.

Figure 3. TRaMs reduce accumulated substrate following xenotransplantation

1. A, B

   Levels of globotriaosylceramide (Gb₃) were measured by LC/MS in plasma and extracts of liver, spleen, heart, and kidneys of mice engrafted with healthy donor (HDo)‐derived (A) or Fabry donor (FDo)‐derived (B) TRaMs, non‐transduced (NT) cells, or in sham‐treated NOD/SCID (NS) and NSF mice.

2. C, D

   Levels of lyso‐Gb₃ were measured by LC/MS in plasma and extracts of tissues of mice engrafted with HDo‐derived (C) or FDo‐derived (D) TRaMs.

Data information: Sphingolipid concentrations are reported as values for individual mice (n = 4) with a line indicating the mean, and error bars showing standard error of the mean. Data from (A–D) were compared using one‐way analyses of variance with Tukey's multiple comparison tests, ***P < 0.001, **P < 0.01, *P < 0.05, and ns P > 0.05. Abbreviations—TRaMs: T‐Rapa micropharmacies; LV/GLA: lentivirus transduced to express α‐gal A; n.d.: not detected. Source data are available online for this figure.

Figure EV5. Shifts in proportions of acyl‐chain variants (ACVs) of Gb₃ after xenotransplantation of TRaMs are small and inconsistent and suggest no species bias of transgenic α‐gal A

1. A, B

   Proportions of Gb₃ ACVs as determined by LC/MS in the indicated tissues in mice transplanted with healthy donor‐derived (A) or Fabry donor‐derived (B) TRaMs compared to controls.

Data information: Percentages reported are means of n = 4 mice, error bars are standard error of the mean. Data were compared using a two‐way analysis of variance with Tukey's multiple comparison test. Significant differences (P < 0.05) between ACV proportions of xenografted compared to sham‐treated mice are indicated with * in the respective bar, and those between mice xenografted with NT or LV/GLA modified T‐Rapa as a red outline. Abbreviations—TRaMs: T‐Rapa micropharmacies; LV/GLA: lentivirus transduced to express α‐galactosidase A.