Original Article T cells isolated from G-CSF-treated multiple myeloma patients are suitable for the generation of BCMA-directed CAR-T cells Anthony M. Battram,1 Aina Oliver-Caldés,1,2 Maria Suárez-Lledó,1,2 Miquel Lozano,3 Miquel Bosch i Crespo,1 Núria Martínez-Cibrián,1,2 Joan Cid,1,3 David F. Moreno,1,2 Luis Gerardo Rodríguez-Lobato,1,2 Alvaro Urbano-Ispizua,1,2,4,5 and Carlos Fernández de Larrea1,2,5 1Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), 08036 Barcelona, Spain; 2Department of Hematology, Amyloidosis and Myeloma Unit, Hospital Clínic of Barcelona, 08036 Barcelona, Spain; 3Apheresis & Cellular Therapy Unit, Department of Hemotherapy & Hemostasis, ICMHO (Institut Clínic de Malalties Hem- atològiques i Oncològiques), Hospital Clínic of Barcelona, University of Barcelona, 08036 Barcelona, Spain; 4Josep Carreras Leukaemia Research Institute, 08036 Barcelona, Spain; 5Department of Haematology, University of Barcelona, 08036 Barcelona, Spain Autologous cell immunotherapy using B cell maturation anti- immunotherapy indicated for the treatment of relapsed and refrac- gen (BCMA)-targeted chimeric antigen receptor (CAR)-T cells tory (R/R) MM that was approved by the US Food and Drug Admin- is an effective novel treatment for multiple myeloma (MM). istration (FDA) and European Medicines Agency (EMA),5 and many This therapy has only been used for relapsed and refractory pa- others are currently being investigated in clinical trials.6,7 tients, at which stage the endogenous T cells used to produce the CAR-T cells are affected by the immunosuppressive nature Presently, most developed CAR-T cells are autologous therapies that of advanced MM and/or side effects of previous therapies. An are made by transducing a CAR molecule into endogenous T cells alternative pool of “fitter” T cells is found in leukocytoaphere- derived from the patient that is destined to receive the treatment.8 sis products that are routinely collected to obtain hematopoiet- For MM, current clinical trials for anti-BCMA CAR-T cells have re- ic progenitor cells for autologous stem cell transplantation cruited patients who have relapsed multiple times in response to (ASCT) early in the treatment of MM. However, to mobilize several therapeutic agents, including the three main families of drugs the progenitor cells, patients are dosed with granulocyte col- for this disease; proteasome inhibitors (bortezomib and carfilzomib), ony-stimulating factor (G-CSF), which is reported to adversely immunomodulatory drugs (lenalidomide and pomalidomide), and affect T cell proliferation, function, and differentiation. Here, anti-CD38 monoclonal antibodies.9 However, at this stage in the dis- we aimed to first establish whether G-CSF treatment negatively ease, T cell abnormalities are promoted by cytokines secreted by ma- influences T cell phenotype and to ascertain whether previous lignant plasma cells and suppressive immune cells as well as other exposure of T cells to G-CSF is deleterious for anti-BCMA thus far uncharacterized mechanisms.10–12 Furthermore, previous CAR-T cells. We observed that G-CSF had a minimal impact lines of therapy can have adverse consequences for T cell immunity, on T cell phenotype when added in vitro or administered to such as dexamethasone-mediated abrogation of T cell proliferation patients. Moreover, we found that CAR-T cell fitness and and depletion of naive CD8+ T cells by the anti-CD38 antibody dar- anti-tumor activity were unaffected when generated from atumumab.13–15 This undesirable combination of disease- and drug- G-CSF-exposed T cells. Overall, we showed that ASCT apher- induced effects results in an immunosuppressive microenvironment esis products are a suitable source of T cells for anti-BCMA and T cell deficiency, which manifests itself as exhaustion, senescence, CAR-T cell manufacture. loss of early-differentiated cells, an alteration in the CD4:CD8 ratio, and metabolic dysfunction.11,14,16–18 Analysis of the parameters that influence the efficacy of one anti-BCMA CAR-T cell product revealed INTRODUCTION Multiple myeloma (MM) is a B cell malignancy that accounts for almost 1% of all newly diagnosed cancers.1 Although treatment op- Received 7 January 2022; accepted 16 June 2022; https://doi.org/10.1016/j.omtm.2022.06.010. tions for MM have undergone radical improvements over recent de- Correspondence: Anthony M. Battram, PhD, Institut d’Investigacions Bio- cades, the disease remains incurable, and relapse is inevitable in mèdiques August Pi i Sunyer (IDIBAPS), C/Rosselló 149-153, 08036 Barcelona, almost all cases.2 Chimeric antigen receptor (CAR)-modified T cells Spain. that target B cell maturation antigen (BCMA) have revolutionized E-mail: battram@clinic.cat Correspondence: Carlos Fernández de Larrea, MD, PhD, Institut d’Investigacions MM therapy and are likely to become a standard part of the treatment Biomèdiques August Pi i Sunyer (IDIBAPS), C/Rosselló 149-153, 08036 Barcelona, regimen for this hematological cancer.3,4 Indeed, the anti-BCMA Spain. CAR-T cell product idecabtagene vicleucel became the first cell-based E-mail: cfernan1@clinic.cat Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 ª 2022 The Authors. 207 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Molecular Therapy: Methods & Clinical Development A B C D Figure 1. ARI2 h expansion and phenotype are unaffected by recombinant G-CSF (A) Schematic of cell culture protocol for production of ARI2h cells with (+G-CSF ARI2) or without (G-CSF ARI2) exposure to 10 ng/mL G-CSF. (B) Cells were stained with annexin V (AnV) and 7-AAD on day 3 (d3) before exposure to G-CSF and afterward on day 0 (d0) to determine the frequency of live cells (AnV7-AAD), early apoptotic cells (AnV+7-AAD), and late apoptotic or dead cells (AnV+7-AAD+). See also Figure S2C. (C) Growth of untransduced (UT) and ARI2h cells was measured as number of population doublings on day 9 compared with day 0. (D) Summary of memory and effector phenotypes of expanded CAR+ ARI2h cells is shown, categorized as early differ- entiated (early, CCR7+CD45RA+), central memory (CM) (CCR7+CD45RA), effector memory (EM) (CCR7CD45RA), and effector memory CD45RA+ (CCR7CD45RA+) cells. Data represent the mean percentage of each population. n = 3. All error bars show mean ± SD. ns, not significant. that some of these factors, particularly fewer CD8+ early differentiated (Treg) expansion.28 Such impairments could reduce the effectiveness (CD45ROCD27+) cells and a lower CD4:CD8 ratio in the starting of CAR-T cells that are generated from G-CSF-exposed T cells, but T cell product, were associated with a poorer clinical response.19 this is currently unknown. Moreover, apheresis products are replete Taken together, these studies clearly indicate that, for CAR-T cell with myeloid cells, including CD14+ monocytes, which are known generation, it would be beneficial to use T cells isolated earlier in to limit CAR-T cell expansion.29,30 the treatment program, before a refractory disease develops. In this report, we studied whether G-CSF adversely affected freshly An alternative and readily available source of “fitter” T cells are the isolated T cells, with particular attention paid to characteristics that apheresis products that are routinely collected and cryopreserved had previously been associated with anti-BCMA CAR-T cell clinical when harvesting progenitor cells for autologous stem cell transplan- responses. Next, anti-BCMA CAR-T cells were generated from tation (ASCT), which is normally conducted early in the treatment T cells previously exposed to G-CSF in vivo and the expanded plan of MM patients.20 To mobilize progenitor cells from the bone CAR-T cells were assessed for phenotype and functionality. The marrow into the peripheral blood ready for collection, patients are BCMA-targeting CAR molecule used in this study is a fully academi- administered with granulocyte colony-stimulating factor (G-CSF).21 cally developed CAR called ARI0002h (henceforth referred to as Either directly or indirectly, G-CSF is proposed to have negative ARI2h),31,32 which is currently being explored in a clinical trial for pa- effects on conventional T cells, such as reducing proliferation,22 im- tients with R/R MM (CARTBCMA-HCP-01 trial, NCT04309981). pairing CD8+ T cell effector function,23,24 causing an imbalance in Overall, we found that the impact of G-CSF was minimal and, most the ratio of Th1/Th2 cells,25–27 and inducing regulatory T cell importantly, did not diminish CAR-T cell activity. 208 Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 www.moleculartherapy.org A ARP-1 U266 125 125 UT UT % surviving MM cells % surviving MM cells 100 -G-CSF ARI2 100 -G-CSF ARI2 +G-CSF ARI2 +G-CSF ARI2 75 75 50 50 25 25 0 0 1:1 0.5:1 0.25:1 0.125:1 1:1 0.5:1 0.25:1 0.125:1 E:T Ratio E:T Ratio B -G-CSF UT +G-CSF UT C -G-CSF UT +G-CSF UT -G-CSF ARI2 +G-CSF ARI2 -G-CSF ARI2 +G-CSF ARI2 ns ns ns * 100 100 10 10 IFNγ (ng/ml) IL-2 (ng/ml) 1 1 0.1 0.1 0.01 0.01 0.01 0.01 0 0 Alone +ARP1 +U266 Alone +ARP1 +U266 D 12 E Control (d0) -G-CSF ARI2 +G-CSF ARI2 ns CD4+ CD8+ Granzyme B (MFI x10 ) 3 9 %maximum 6 ns 3 0 e e 1 1 P- P- on on AR AR al al SF SF + + -C -C CD25-PE/Cy7 SF SF -G +G -C -C -G +G F CD4+ CD8+ -G-CSF ARI2 +G-CSF ARI2 -G-CSF ARI2 +G-CSF ARI2 250 250 Normalised MFI (% PRE UT) Normalised MFI (% PRE UT) ns ns ns ns 200 200 ns ns 150 150 ns * 100 100 50 50 0 0 PD-1 LAG-3 TIM-3 TIGIT PD-1 LAG-3 TIM-3 TIGIT Figure 2. Recombinant G-CSF does not diminish ARI2 h activity or promote ARI2 h exhaustion (A) Survival of GFP-ffLuc-expressing ARP-1 (left) or U266 (right) myeloma cell lines following a 24-h co-culture with UT or ARI2h cells at the indicated T cell:tumor cell line (effector:target [E:T]) ratios, determined by measuring target cell bioluminescence (n = 4). Data are represented as mean ± SEM. (B–D) UT and ARI2h cells were co-incubated (legend continued on next page) Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 209 Molecular Therapy: Methods & Clinical Development RESULTS Patient G-CSF treatment did not alter the T cell CD4:CD8 ratio, Recombinant G-CSF had no direct effect on ARI2 h expansion or increase the proportion of Tregs or induce T cell dysfunction activity After establishing that G-CSF did not directly impact ARI2h cells Although many of the inhibitory actions of G-CSF on T cells are sec- in vitro, we explored the possibility of generating ARI2h cells from pa- ondary effects caused by the regulation of other cell types, it has been tients who received G-CSF to mobilize CD34+ progenitor cells prior reported that G-CSF directly controls T cell function.25,33 To examine to ASCT. All patients were diagnosed with MM, and ASCT was being whether these direct effects impact anti-BCMA CAR-T cell activity, used as part of a first-line therapy strategy following induction treat- T cells were cultured for 3 days in the presence of recombinant ment (Table S1). The induction therapy used for all patients was G-CSF prior to activation, ARI2h CAR transduction, and cell expan- similar both in terms of drugs administered and number of treatment sion (Figure 1A). The efficacy of the recombinant G-CSF was tested cycles (Table S2). by analyzing G-CSF-induced STAT3 phosphorylation in monocytes (Figure S1).34 There was no loss in T cell viability following the To investigate the impact of G-CSF on CAR-T cells generated from initial pre-stimulation incubation stage, and importantly, the MM patient T cells, first it was necessary to study the potential ef- addition of G-CSF did not impair T cell survival or induce apoptosis fects of G-CSF on T cells before CAR-T cell generation. As such, (Figures 1B and S2). At the end of the ARI2h production, it was blood was collected from patients before (“PRE”) and after found that pre-incubation with G-CSF had no effect on cell growth (“POST”) 4 days of G-CSF treatment (10–12 mg/kg every 12 h) in terms of either population doublings or differentiation and subject to analysis of T cell populations and phenotype (Fig- (Figures 1C and 1D). ure 3A). During the T cell isolation step, monocytes were carefully removed to avoid potential issues with CAR-T cell expansion (Fig- When ARI2h cells were challenged with two MM cell lines, ure S3).29 From the isolated T cells, the relative proportion of CD4+ ARP-1 and U266, it was observed that G-CSF had no impact on and CD8+ T cells varied dramatically (0.93–3.16), as previously the anti-tumor killing ability of the CAR-T cells (Figure 2A). described for MM patients,18 but there was no significant difference Correspondingly, G-CSF exposure had no bearing on tumor-cell- in the CD4:CD8 ratio between PRE and POST samples (p = 0.1595) induced interferon g (IFNg) secretion (Figure 2B). Although (Figure 3B). As G-CSF has been described to cause Treg expan- there was a small difference in interleukin-2 (IL-2) release in sion,28 the frequency of Tregs, characterized as CD25highFoxP3+ response to ARP-1 (p = 0.047), the average reduction in IL-2 cells, was assessed. Tregs made up less than 7% of the total CD4+ was small (6.6 ng/mL) in comparison to the total amount found T cells in all analyses, with no difference observed between PRE in the supernatant, and no difference was observed with U266 cells and POST samples (p = 0.794) (Figure 3C). (Figure 2C). Similarly, G-CSF did not alter ARI2h granzyme B levels, either basally or following stimulation with tumor cells Another potential factor that could influence the quality of CAR-T (Figure 2D). cells is the dysfunctionality of the T cells in the starting material. To compare the exhaustion phenotype between PRE and POST Another aspect of CAR-T cells that is thought to contribute to CD8+ T cells, expression of the checkpoint molecules PD-1, LAG-3, in vivo functionality is their expression of key activation and inhib- and TIGIT were analyzed, but levels of these markers were similar itory molecules. After expansion, ARI2h CD4+ and CD8+ cells ex- (Figure 3D). TIM-3 expression was also evaluated but was too low pressed high levels of CD25, the IL-2 receptor a chain (Figure 2E), to make accurate comparisons (data not shown). Another feature of which is used to measure CAR-T cell activation through IL-2 T cell fitness is senescence, which can be characterized as the presence signaling but can also demonstrate tonic signaling when above of CD27CD28 cells within the CD8+ pool.36 In some patients, se- normal levels.35 G-CSF did not alter ARI2h CD25 expression, sug- nescent-like cells constituted more than 40% of CD8+ T cells, but this gesting that G-CSF neither decreased cell activation nor induced was consistent in both PRE and POST G-CSF (Figure 3E). spontaneous tonic signaling. In addition, levels of checkpoint mol- ecules PD-1, LAG-3, TIM-3, and TIGIT were not augmented by In vivo G-CSF reduced frequency of TSCM cells but had no effect G-CSF; indeed, TIM-3 expression was reduced on CD8+ cells on other effector and memory T cell populations (Figure 2F). The effector and memory phenotypes of T cells are another important determinant of their activity, and for anti-BCMA CAR-T cells, the Collectively, these data demonstrate that G-CSF did not exert any presence of early differentiated memory cells within the starting direct in vitro effects on T cells that subsequently impaired CAR-T T cell product correlates with patient outcomes.19 When comparing cell proliferation, function, or phenotype. the frequency of effector and memory cell populations between with ARP-1 cells, U266 cells, or media alone (alone) for 6 h. Levels of IFNg (B) and IL-2 (C) released into the supernatant were evaluated by ELISA, and the expression of granzyme B in CAR+ cells was assessed by flow cytometry, displayed as median fluorescence intensity (MFI) (D). Error bars show mean ± SD. (E) Histogram of CD25 expres- sion on CAR+ CD4+ (left) and CD8+ (right) T cells and unstimulated day 0 T cells (control), representative of three experiments, is shown. (F) Summary of the surface expression of PD-1, LAG-3, TIM-3, and TIGIT on CAR+ CD4+ (left) and CD8+ (right) T cells is shown, represented as MFI normalized to that displayed on UT PRE G-CSF UT cells. Error bars show mean ± SD.*p < 0.05; ns, not significant. 210 Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 www.moleculartherapy.org A B 4 10-12 μg/kg ns G-CSF 3 Patient Day: 0 1 2 3 4 CD4:CD8 ratio 'PRE' 'POST' 2 PBMCs PBMCs T-cell isolation 1 'PRE' 'POST' 0 PRE POST C PRE POST 8 ns 6 + CD25-PE/Cy7 FoxP3 high 4 5.15% 5.99% %CD25 2 0 FoxP3-PE PRE POST D ns ns 1000 ns 100 500 800 400 75 LAG-3 MFI TIGIT MFI PD-1 MFI 600 300 50 400 200 25 200 100 0 0 0 PRE POST PRE POST PRE POST E PRE POST ns 50 40 16.1% 18.4% CD28-FITC - % CD27 CD28 30 - 20 10 0 CD27-PE PRE POST (legend on next page) Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 211 Molecular Therapy: Methods & Clinical Development PRE and POST CD4+ and CD8+ T cells, based on CD45RA and CCR7 higher percentage of CAR+ cells compared with the corresponding staining, no difference was observed (Figure 4A). To further identify POST ARI2h cultures in six out of seven expansions (p = 0.006) (Fig- the presence of distinct memory subsets, the expression of CD95 was ure 5E). However, it is worth noting that the number of T cells ob- evaluated on “early” differentiated CD45RA+CCR7+ cells, which dis- tained from all patients after 4 days of G-CSF treatment was higher tinguishes between naive populations and those with sustained self- than before they had been given G-CSF (p = 0.008) (Figure 5F). renewal potential, termed stem-cell-memory-like (TSCM) cells.37 This 2-fold increase in isolated T cells allows for the generation of Although the proportion of CD4+ and CD8+ naive (CD95) cells more ARI2h cells from a given patient, despite the 22% reduction was unchanged by G-CSF exposure (Figure 4B), the frequency of in transduction efficiency in POST ARI2h cultures. CD8+ TSCM (CD95+) cells was 40% lower in POST G-CSF samples (p = 0.008) (Figure 4C). In summary, POST ARI2h cells expanded as well as PRE ARI2h cells and, although there was a difference in transduction rates, this was An alternative approach to identify T cell memory phenotypes is to stain offset by the fact that many more T cells are obtained from the for CD45RA/CD45RO and CD27, which is particularly relevant to same volume of blood following G-CSF administration. CAR-T cell production, as the presence of CD8+CD45ROCD27+ cells in the starting material is a known predictor of patient outcomes.19,38 Analysis of CD8+CD45RA+CD27+ cells, which are equivalent to ARI2h cells generated from PRE and POST T cells demonstrate CD8+CD45ROCD27+ cells,39 revealed that the proportion of this equal anti-tumor activity in vitro and in vivo clinically important population was unaltered by G-CSF (Figure 4D). To determine whether G-CSF affected POST ARI2h cell functional ability, PRE and POST ARI2h cells were challenged with MM tumor Put together, the data from T cells before and after G-CSF adminis- cell lines and their responses were analyzed. Both PRE and POST tration demonstrate that G-CSF had only a minor impact on T cell ARI2h cells exerted a similar cytotoxicity against MM cells and phenotype and populations, with the only identified change being secreted equal levels of effector molecules, including granzyme B, the proportion of the small pool of TSCM cells. IFNg, and IL-2 (Figures 6A–6D and S4A). Furthermore, the prolifer- ative responses of PRE and POST ARI2h cells to MM cell line ARP-1 were equivalent (Figures 6E and S4B). High expansion rates and increased T cell isolation after G-CSF treatment compensated for reduced CAR transduction To test the persistence of PRE and POST ARI2h cell responses, they After establishing that T cells from patients after G-CSF treatment were repeatedly stimulated with ARP-1 cells in vitro and subsequently were suitable for CAR-T cell production, ARI2h cells were expanded assessed for their ability to expand and kill the tumor cells (Fig- as shown in Figure 5A. Sufficient proliferation of ARI2h cells was ure S5A). For a period of four consecutive challenges, PRE and observed regardless of whether the T cells had been exposed to POST ARI2h cells exhibited similar large-scale expansion (average G-CSF, with population doubling rates exceeding four for 75% of doubling rates of 10.58 and 10.51 for PRE ARI2h and POST the PRE and POST cultures (Figure 5B). Interestingly, untransduced ARI2h, respectively) and rapid tumor cell killing that was maintained (UT) cells expanded from POST G-CSF T cells actually grew better with each stimulation (Figures 6F, 6G, and S5B). After the repeated than UT cells from PRE G-CSF T cells. The viability of UT or stimulations, almost all T cells in the PRE and POST ARI2h cell ARI2h cells was 85% or higher (Figure 5C), with the exception of cultures were CAR+ (Figure S5C), with a similar differentiation status one PRE ARI2h culture, which was still above the limit (70%) for and expression of exhaustion-associated inhibitory receptors satisfactory clinical production of ARI2h.31 Other parameters that (Figures S5D–S5F). are often important for CAR-T cell production are the CD4:CD8 ratio and the CAR transduction rate. In the majority of both the PRE and To better examine the relative efficacies of PRE and POST ARI2h POST ARI2h cultures, CD4+ T cells outnumbered CD8+ T cells on cells, they were tested in a xenograft mouse model of MM with stress day 7 of the expansion, but the CD4:CD8 ratio showed a high degree conditions to severely challenge anti-BCMA CAR-T cells. For this of variability between donors (0.7–9.9) (Figure 5D). experiment, irradiated mice were engrafted with firefly luciferase (ffLuc)-expressing ARP-1 cells until they developed a substantial tu- All CAR-T cell cultures achieved a CAR transduction efficiency that is mor burden, after which they were dosed with PRE or POST ARI2h suitable for clinical purposes,31 but PRE ARI2h cultures exhibited a cells or UT T cells as a control (Figure 7A). Whereas the animals given Figure 3. Patient populations of Tregs and dysfunctional T cells are not increased following G-CSF administration (A) Schematic of G-CSF dosing. Patients were administered with 10–12 mg/mL G-CSF every 12 h for a total of 4 days. Before (PRE) and after (POST) G-CSF treatment, PBMCs were harvested from the patients, from which T cells were isolated. (B) Ratio of CD4+ and CD8+ cells in freshly isolated T cells is shown. (C) Frequency of CD4+ CD25highFoxP3+ Tregs within the PRE G-CSF and POST G-CSF samples is shown as FACS plots (left) and summarized as a percentage of total CD4+ T cells (right). Numbers on FACS plots indicate the percentage of cells found within the Treg gate. (D) Expression of exhaustion markers PD-1 (left), LAG-3 (middle), and TIGIT (right) on PRE G-CSF and POST G-CSF CD8+ T cells is shown, displayed as MFI. (E) Senescent-like CD27CD28 cells were identified by flow cytometry and are presented as FACS plots (left) and summarized as a percentage of total CD8+ T cells (right). Numbers on FACS plots indicate the percentage of cells found within the CD27CD28 gate. Data points in (B)–(E) are coded according to patient donor and are consistent with other figures. ns, not significant. 212 Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 www.moleculartherapy.org A CD4+ EMRA EMRA PRE POST CCR7-PerCP/Cy5.5 EM EM Early Early CM CM PRE POST CD45RA-APC CD8+ PRE POST CCR7-PerCP/Cy5.5 EMRA Early EMRA Early CM CM EM EM PRE POST CD45RA-APC B C + + + + CD4 CD8 CD4 CD8 80 60 ns 8 4 ** ns ns 60 6 3 40 % naïve % naïve % TSCM % TSCM 40 4 2 20 20 2 1 0 0 0 0 PRE POST PRE POST PRE POST PRE POST D + CD8 PRE POST 80 ns CD45RA-APC 60 + % CD45RA CD27 + 40 20 0 CD27-PE PRE POST Figure 4. Presence of stem cell memory T cells diminished by G-CSF, but other memory and effector populations are unchanged (A) Memory and effector phenotypes of patient-derived CD4+ (upper row) and CD8+ (lower row) T cells before (PRE) and after (POST) G-CSF administration, categorized as early differentiated (early) (CCR7+CD45RA+), CM (CCR7+CD45RA), EM (CCR7CD45RA), and effector memory CD45RA+ (CCR7CD45RA+) cells. Summarized in pie (legend continued on next page) Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 213 Molecular Therapy: Methods & Clinical Development PRE or POST ARI2h cells exhibited a decrease or stabilization in dis- T cells that had previously been exposed to G-CSF. This included ease severity upon T cell administration, those injected with UT cells no augmentation in the expression of checkpoint molecules, alter- rapidly developed terminal levels of cancer (Figure 7B, day 33). Anal- ation in memory and effector populations, or loss of efficacy, either ysis of tumor-cell bioluminescence in the following weeks demon- in vitro or in a murine xenograft model. strated that the disease developed similarly in mice treated with PRE or POST ARI2h cells (Figures 7B, 7C, and S6). In turn, there Before analyzing CAR-T cells, we examined the effects of G-CSF on was no difference in either the survival time of animals from the T cells to ensure that they were suitable for CAR-T cell production. PRE or POST groups (p = 0.990) (Figure 7D) or the persistence of Incubation of T cells with recombinant G-CSF excluded most direct PRE or POST ARI2h cells (Figure 7E). influences of G-CSF on T cell phenotype and CAR-T cell function, although it did cause a reduction in ARI2h IL-2 production and Overall, these data show that the activity of ARI2h cells is unaffected TIM-3 expression on CD8+ cells. Moreover, we cannot rule out a by previous exposure to G-CSF. more general role of G-CSF on T cell function, as this study only eval- uated the anti-tumor activity of CAR-T cells in response to MM cells. G-CSF treatment does not negatively impact the ex vivo Indeed, a direct role of G-CSF on T cell effector function has been phenotype of expanded ARI2h cells demonstrated in the context of anti-viral responses.23,24 To further explore how G-CSF impacts parameters that could influence the activity of ARI2h cells in patients, the phenotypes of PRE and POST It was important to study the effect of G-CSF in vivo, as well as ARI2h cells were analyzed. First, the differentiation status of ARI2h in vitro, because G-CSF-mediated regulation of other immune cell cells from PRE and POST cultures was determined by CD45RA and types, such as dendritic cells,40 monocytes,41 and myeloid-derived CCR7 staining. The proportion of each memory and effector cell suppressor cells (MDSCs),42 also causes T cell suppression.33 For type was alike, with the only exception being the 20% reduction in example, increased MDSC mobilization to the peripheral blood the CD4+ central memory pool (p = 0.018) (Figure 8A). Likewise, promotes abnormalities in the T cell compartment, such as reduced CXCR3 staining of early differentiated CD45RA+CCR7+ cells, which proliferation and polarization of CD4+ cells from a Th1 phenotype indicates the presence of TSCM-like cells,37 was not altered (Figure 8B). towards a Th2 phenotype.43,44 Moreover, the increased presence Analysis of activation markers demonstrated that CD25 was of histocompatibility leukocyte antigen DR isotype (HLA- expressed at similar levels on PRE and POST ARI2 h cells (Figure 8C), DR)/lowCD33+CD16 MDSCs in peripheral blood causes T reg but CD95 was significantly reduced on both CD4+ and CD8+ ARI2h expansion,44 highlighting the complex interplay between G-CSF cells within the POST culture (p = 0.003 and p = 0.011, respectively) and a wide range of immune cells. From our results, however, we (Figure 8D). Similar to CD95, expression of the checkpoint molecule did not observe any negative effects on T cell phenotype or prolifer- LAG-3 was lower on POST ARI2h cells (p = 0.034 and p = 0.035 on ation that could have occurred as a result of the indirect action of CD4+ and CD8+ cells, respectively), but other exhaustion markers G-CSF on T cells via MDSCs or other cell types. Indeed, T cell pro- PD-1, TIM-3, and TIGIT were unchanged (Figures 8E and S7). liferative responses and patient T reg numbers were normal after G-CSF infusion. This discrepancy with previously published data Together, these results show that the phenotype of POST ARI2h cells could be due to the effects of MM or induction therapy on the im- is suitable for therapeutic use and, moreover, confers some potential mune system, thus dampening any potential effect of G-CSF, or advantages, such as lower LAG-3 expression. possibly because the POST G-CSF samples were collected immedi- ately after G-CSF administration following only a short exposition DISCUSSION to the drug, meaning that any effect of G-CSF had not yet material- In this study, the functionality of ARI2h cells that had been exposed to ized. Supporting this hypothesis, Toh et al.45 showed that, during G-CSF prior to ex vivo expansion was analyzed, with the aim of deter- the G-CSF dosage period, Treg frequency was the same as in control mining whether these T cells could be used as a starting material for samples, but 1–3 weeks later, the percentage of Tregs was significantly the production of CAR-T cells. After demonstrating that G-CSF had higher. Another important difference between our study and that of no impact on the activity of ARI2h cells when added in vitro, we some other groups is that we used highly purified T cells as a starting examined the effect of G-CSF on patient T cells and the ARI2h cells material for the ex vivo culture of the CAR-T cells and not peripheral derived from them. We observed that, in this context, G-CSF did not blood mononuclear cells (PBMCs). This approach matches the clin- alter T cell characteristics or populations, with the exception of ical protocol used at our institution31,46 and confers the major advan- reducing the TSCM pool. However, despite this potential drawback, tage of removing T cell inhibitory cell types, such as CD14+ mono- ARI2h cells displayed no signs of inferiority when produced from cytes, that severely hinder CAR-T cell expansion.29,30 charts (left) shows the mean percentage of each population and displayed as FACS plots. n = 4. (B and C) Measurement of CD95 expression on early differentiated CD4+ (left) and CD8+ (right) CCR7+CD45RA+ cells enabled the identification of CD95 naive cells (B) and CD95+ stem cell memory T (TSCM) cells (C). Data are shown as a percentage of total CD4+ or CD8+ cells. (D) CCR7/CD45RA staining of patient T cells to identify the CCR7+CD45RA+ early differentiated population is shown. Data are displayed as example FACS plots (left) and summary of percentage of CCR7+CD45RA+ cells from all donors (right). Data points in (B)–(D) are coded according to patient donor and are consistent with other figures. **p < 0.01; ns, not significant. 214 Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 www.moleculartherapy.org A B C D E F Figure 5. Decrease in CAR transduction of POST G-CSF T cells is compensated by an increase in the concentration of isolated T cells (A) Schematic of cell culture protocol for production of ARI2h cells from patient-derived T cells that were from before (PRE ARI2) or after (POST ARI2) G-CSF administration. (B) Growth of untransduced (UT) and ARI2 h cells measured as number of population doublings on day 9 compared with day 0 is shown. (C) UT and ARI2 h cells were stained (legend continued on next page) Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 215 Molecular Therapy: Methods & Clinical Development From all three factors that have been linked with the in vivo effective- mobilized apheresis products collected for ASCT will be compared ness of CAR-T-BCMA cells and patient responses, which are (1) the with peripheral blood drawn from R/R MM patients (i.e., current clin- frequency of CD8+ early differentiated cells, (2) the CD4:CD8 ratio ical practice) as a starting material for CAR-T cell production. of the pre-manufactured T cells, and (3) the in vitro expansion rate of the CAR-T cells,19 all were unaltered by G-CSF. Early differentiated Overall, investigations to improve the starting material for anti- cells include both naive memory cells and TSCM cells, both of which are BCMA CAR-T cell production, such as this study, are timely due to believed to have unique roles in the durability of T-cell-based immu- the substantial number of anti-BCMA CAR-T clinical trials that are notherapies.47 Here, we showed that, while the number of CD8+ naive ongoing and due to the relevance of our results to the production cells (CD45RA+CCR7+CD95 or CD45RA+CD27+) was unchanged of all types of CAR-T cells. In fact, an additional advantage to using by G-CSF, CD8+ TSCM cells (CD45RA+CCR7+CD95) were dimin- apheresis products collected for ASCT would be to decrease the ished. Although the proportion of TSCM in the final CAR-T cell prod- burden of apheresis units that are now highly saturated since they uct is correlated with patient responses,48 it is currently unclear must cater to both ASCT and the collection of material for CAR-T whether the proportion of TSCM cells in unstimulated T cells influ- cell manufacture. To conclude, we believe that we have provided ences anti-BCMA CAR-T cell therapy. However, ARI2h cells made strong evidence to support the incorporation of apheresis-derived from G-CSF-treated T cells had equal in vivo function to those not T cells as a starting material for CAR-T cell production in the clinic. exposed to G-CSF, suggesting that the level of reduction of the TSCM pool caused by G-CSF is insignificant for anti-tumor efficacy. What MATERIALS AND METHODS is more, CXCR3 expression in the CD45RA+CCR7+ population was Human samples the same in PRE and POST ARI2h cells, which indicates similar fre- MM patients who were in line to receive ASCT were given recombi- quencies of TSCM-like cells in the final CAR-T cell product. Another nant G-CSF at a dose of 10–12 mg/kg/12 h for a total of 4 days to notable property of the infusion CAR-T cell product that was also un- facilitate the collection of CD34+ progenitor cells by apheresis. Imme- affected by patient G-CSF administration is the CD4:CD8 ratio, which diately before the first G-CSF dose and shortly after the final G-CSF was found to be important for the efficacy of some CAR-T cell administration just before the apheresis collection, blood was drawn products,49,50 but not others.19,38 from patients into EDTA-coated tubes. Patients did not receive pler- ixafor before blood collection. Buffy coats from healthy donors were Another T cell intrinsic quality that is believed to affect CAR-T cell obtained from the local blood and tissue bank (Banc de Sang i Teixits, efficacy is dysfunction. In fact, studies of CD19-directed CAR-T cells Catalonia). T cells were isolated from fresh blood or buffy coats by found that the exhaustion profile of pre-infusion CAR-T cells also density-gradient centrifugation using Histopaque-1077 (Sigma-Al- correlated with patient responses.38,51 For ARI2h, lower expression drich, St. Louis, MO) followed by negative selection of T cells using of checkpoint inhibitory molecules appears to correlate with activity a Pan T Cell Isolation kit (Miltenyi Biotech, Bergisch Gladbach, Ger- as well.52,53 Our data revealed that the expression of the checkpoint many). Monocytes were purified from buffy coats using the receptor LAG-3 was lower in POST ARI2h CD8+ cells compared RosetteSep Human Monocyte Enrichment Cocktail (STEMCELL with PRE ARI2h CD8+ cells, and importantly, the levels of other Technologies, Vancouver, Canada). exhaustion markers (PD-1, TIM-3, and TIGIT) were not increased. LAG-3 expression on T cells appears to be particularly important Ethical approval for MM, because it is associated with event-free survival.54 Indeed, Informed written consent in compliance with the Declaration of Hel- in our previous study, we discovered that the IL-15-cultured ARI2h sinki was obtained from all donors. Research involving human- cells that performed better in vivo also had lower levels of LAG-3 prior derived material and mice was approved by the Ethical Committees to infusion,52 suggesting a link between ARI2h activity and LAG-3 of Clinical Research (Hospital Clínic, Barcelona) and Animal expression. Research (University of Barcelona, Barcelona), respectively. The next stage to ensure the clinical translation of this study would be Cell culture and T cell transduction to validate the use of frozen and stored apheresis products as a viable T cells were stimulated with Dynabeads Human T-Activator CD3/ option for CAR-T cell generation instead of fresh blood, as was used CD28 (Thermo Fisher Scientific, San Diego, CA) and were expanded in this study. Moreover, further investigation is required into the in T cell media (47.5% Click’s media [Irving Scientific, Santa Ana, intrinsic qualities that cause T cells from patients early in their MM CA], 47.5% RPMI-1640, 5% human serum, 2 mM L-glutamine, 100 evolution to be more favorable for CAR-T cell production than those IU/mL penicillin, and 100 mg/mL streptomycin) supplemented with from R/R and heavily treated patients. We plan to address these issues 100 IU/mL IL-2. Forty-eight hours following stimulation, T cells in a novel line of investigation in which the use of frozen and stored were transduced with a lentivirus vector encoding ARI2h.31 with annexin V (AnV) and 7-AAD to assess viability, presented as percentage of live cells (AnV-7-AAD). (D) Ratio of CD4+ and CD8+ cells in UT and CAR+ ARI2 h cells on day 9 is shown. (E) Percentage of PRE ARI2 and POST ARI2 cells that were positive for the ARI2 h CAR molecule on day 7 is shown. (F) Number of T cells isolated from the blood of patients before (PRE) and after (POST) G-CSF treatment is shown, normalized to the volume of blood drawn. All error bars show mean ± SD, and data points in (B)–(F) are coded according to patient donor and are consistent with other figures. **p < 0.01; ns, not significant. 216 Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 www.moleculartherapy.org A UT B ARP-1 U266 10 ns PRE ARI2 125 POST ARI2 125 8 Granzyme B (MFI x10 ) 3 % surviving MM cells % surviving MM cells 100 100 6 75 75 4 ns 50 50 2 25 25 0 0 0 2:1 1:1 0.5:1 0.25:1 2:1 1:1 0.5:1 0.25:1 e e 1 1 P- P- on on AR AR al al ST E:T Ratio E:T Ratio + + E PR PO E ST PR PO C D CAR a lone +ARP-1 +U266 CAR a lone +ARP-1 +U266 ns ns ns ns 100 100 10 10 IFNγ (ng/ml) IL-2 (ng/ml) 1 1 0.1 0.1 0.01 0.01 0.01 0.01 0 0 E T E T E T E T E T E T ST RI2 ST RI2 ST RI2 ST RI2 ST RI2 ST I2 I2 I2 I2 I2 I2 I2 PR U PR U PR U PR U PR U U R R R R R R R PO A PO A PO A PO A PO A PO A A A A A A A E F PR 12 PRE ARI2 POST ARI2 Population doublings 9 6 POST +ARP-1 %maximum PRE +ARP-1 3 PRE UT POST UT POST alone 0 Before C1 C2 C3 C4 PRE alone G 100 PRE ARI2 % surviving MM cells Unlabelled POST ARI2 75 CFSE 50 25 0 0 24 48 72 Hours post challenge (legend on next page) Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 217 Molecular Therapy: Methods & Clinical Development Subsequently, T cells were split and cytokines were refreshed every 1 CD4+CD25highFoxP3+ cells with conventional CD4+ T cells used as to 2 days for a further 7 days of culture, unless indicated otherwise. a negative control (full gating strategy shown in Figure S8). For experiments in which T cells were exposed to G-CSF in culture, 10 ng/mL recombinant G-CSF (Miltenyi Biotech) was added to Analysis of granzyme B expression T cells that were resting in T cell media for 3 days prior to stimulation. Equal numbers of ARI2h and ARP-1 cells were co-cultured for 6 h at ARP-1, U266, ARP-1-GFP-ffLuc, and U266-GFP-ffLuc cell lines were 37 C, with GolgiPlug (BD Biosciences) added at a dilution of 1:1,000 obtained, cultured, and modified to express GFP-ffLuc, where appro- after the 1st hour. The cells were then stained for surface markers, priate, as previously described.31 All cultured cells were incubated at fixed, and permeabilized according to the instructions of the FIX&- 37 C with 5% CO2. Live cells were routinely counted using Trypan PERM Cell Fixation and Permeabilization Kit (Nordic MU Bio, Sus- blue exclusion. teren, the Netherlands) with the inclusion of an AF647-conjugated anti-granzyme B antibody (BD Biosciences) during the permeabiliza- Flow cytometry tion stage. Cell surface proteins on (CAR-)T cells were stained with the following antibodies: annexin V-PE, CCR7-PerCP/Cy5.5, CD3-allophycocya- In vitro cytotoxicity assay nin (APC), CD3-BV421, CD4-APC/H7, CD8a-PE/Cy7, CD25-PE/ To assess ARI2h anti-tumor cell activity, an in vitro luminescence- Cy7, CD69-fluorescein isothiocyanate (FITC), CD95-PE, CD127- based assay was employed, as previously described.50 In brief, PerCP/Cy5.5, CXCR3-AF488 (all from BD Biosciences, San Jose, GFP-ffLuc-expressing ARP-1 (ARP-1-GFP-ffLuc) or U266 (U266- CA); CD8a-APC, CD8a-FITC, CD27-PE, CD28-FITC, CD45RA- GFP-ffLuc) cells were co-cultured with ARI2h cells at the indicated APC, LAG-3-PE, TIGIT-PerCP/Cy5.5, TIM-3-FITC (all from effector:target ratios. D-luciferin was added to the cells 10 min prior BioLegend, San Diego, CA); and PD-1-APC (eBioscience, Thermo to analysis, and the luminescence signal was read on a plate reader. Fisher Scientific). Monocytes were stained with CD14-PE (eBio- ARI2h activity was interpreted based on the luminescence signal science). To determine CAR expression, cells were labeled with a re- emitted by the surviving tumor cells, which was calculated as combinant Fc-tagged BCMA protein (Enzo Life Sciences, Farming- follows: dale, NY) followed by a BV421-conjugated anti-Fc antibody (BioLegend). Samples prepared from mouse organs were treated with Fc block (BD Biosciences) prior to staining. Cells were washed and resuspended in fluorescence-activated cell sorting (FACS) buffer 100  ðsignal fromðsignal from sample well  background signalÞ well containing tumour cells alone  background signalÞ (1% FCS in PBS) or 1% paraformaldehyde prior to acquisition. In ex- periments to assess viability and/or apoptosis, cells were resuspended ELISAs in FACS buffer containing 1 mg/mL 7-aminoactinomycin D (7-AAD). ARI2h cells and ARP-1 or U266 cells were co-cultured for 6 h at a 1:1 In all other experiments, live cells were gated based on forward and ratio, after which the supernatants were collected and stored at side scatter. 80 C until analysis. The concentrations of IFNg and IL-2 in the su- pernatants were analyzed by ELISA using commercially available kits All flow cytometry data were acquired on a FACSCanto II (BD Bio- (BioLegend). sciences) and analyzed using FlowJo software version X.0.7 (Treestar, Ashland, OR). In vivo murine experiments Mouse experiments were performed similarly to those described Staining of Tregs previously.52 In detail, on day 0, 1.5  106 ARP-1-GFP-ffLuc cells Following cell surface staining, T cells were fixed and permeabilized were injected into the tail vein of irradiated 9-week-old female using the Foxp3/Transcription Factor Staining Buffer Set (eBio- non-obese diabetic (NOD)-severe combined immunodeficiency science) according to the manufacturer’s instructions. During the (SCID)-IL-2gc/ mice, and on day 20 post-tumor injection, the permeabilization step, Tregs were stained with a PE-conjugated mice received 8.6  106 UT T cells or 6  106 CAR+ PRE or FoxP3 antibody (BD Biosciences). Tregs were identified as POST ARI2h cells (8.6  106 total cells) through intravenous Figure 6. POST G-CSF ARI2h anti-myeloma cell responses are effective (A) Survival of GFP-ffLuc-expressing ARP-1 (ARP-1-GFP-ffLuc) (left) or U266 (right) myeloma cell lines following a 24-h co-culture with UT or ARI2h cells at the indicated T cell:tumor cell line (effector:target [E:T]) ratios, determined by measuring target cell bioluminescence (n = 4 to 5). Data are represented as mean ± SEM. (B–D) UT and ARI2 h cells were co-incubated with ARP-1 cells, U266 cells, or media alone (CAR alone) for 6 h. The expression of granzyme B in CD8+ CAR+ cells was assessed by flow cytometry (B), and levels of IFNg (C) and IL-2 (D) released into the supernatant were evaluated by ELISA. Error bars show mean ± SD. Data points are coded according to patient donor and are consistent with other figures. See also Figure S4A. (E) ARI2h cells were labeled with CFSE and cultured with (+ARP-1) or without (alone) ARP-1 cells. After 72 h, ARI2 h cell proliferation was analyzed by measuring CFSE dilution. ARI2h cells that were not treated with CFSE (unlabeled) acted as a control. The displayed histograms are CD8+ CAR+ cells and are representative of three experiments. See also Figure S4B. (F and G) Fold expansion of ARI2h cells following four challenges (C1, C2, C3, and C4) with ARP-1-GFP-ffLuc cells at a 0.125:1 E:T ratio (n = 3; F) is shown. Survival of ARP-1-GFP-ffLuc cells during the fourth stimulation is shown, determined by measuring target cell bioluminescence (n = 3) (G). Data are represented as mean ± SEM. ns, not significant. See also Figure S5. 218 Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 www.moleculartherapy.org A Day C 10 9 -1 0 20 28 Radiance (p/s/cm /sr) 2 10 8 10 7 UT ARP-1 UT or UT or 10 6 PRE ARI2 PRE ARI2 or PRE ARI2 or POST ARI2 POST ARI2 POST ARI2 10 5 B 10 20 30 40 50 60 70 Days (post-tumour injection) D 100 Percent survival 75 p=0.039 p=0.015 50 UT * 25 * PRE ARI2 p=0.990 POST ARI2 0 0 20 40 60 Days (post-tumour injection) E Spleen BM 1.0 1.0 ns 0.8 0.8 0.6 0.6 + + %CD3 %CD3 ns 0.4 0.4 0.2 0.2 0.0 0.0 I2 I2 I2 I2 AR AR AR AR E E ST ST PR PR PO PO Figure 7. PRE and POST G-CSF ARI2h cells are similarly effective in a mouse MM model (A) Schematic of in vivo experimental design. On day 0, irradiated mice were injected with 1.5  106 GFP-ffLuc-expressing ARP-1 cells. Following tumor engraftment, mice were transfused twice with UT T cells, PRE G-CSF ARI2h cells, or POST G-CSF ARI2h cells, on day 20 and on day 28. (B) Tumor progression was monitored by weekly readings of animal bioluminescence. Shown are photos taken from the ventral side of the mice. Numbers on the left of the image show the number of days following tumor cell infusion. See also Figure S6C. (C) Quantification of bioluminescence is shown, showing data from each mouse. The dotted horizontal line indicates the predetermined threshold for mouse sacrifice due to heavy tumor burden. (D) Overall survival of mice from each group is shown. (E) Frequency of anti-human CD3+ ARI2h cells in the spleen and bone marrow (BM) of mice immediately following sacrifice is shown. Error bars show mean ± SD. Arrows in (C) and (D) indicate days on which T cells were injected. Data points in (E) are coded according to patient donor and are consistent with other figures. *p < 0.05; ns, not significant. inoculation. Eight days later, on day 28, the mice received a second the investigator analyzing tumor burden and the technician inject- dose of CAR-T cells: 2.9  106 UT T cells or 2  106 CAR+ PRE or ing the T cells into the mice were unaware which mouse received POST ARI2h cells (2.9  106 total cells). Every week, beginning on which T cell treatment. day 7, tumor growth in each mouse was monitored by injection with D-luciferin and measurement of the bioluminescence signal on both the ventral and dorsal sides. The signals were averaged, and when Proliferation measurements the mean bioluminescence exceeded 3  108 p/s/cm2/sr, mice ARI2h cells were starved of IL-2 for 24 h and then stained with car- were sacrificed. The spleen and the bone marrow from the hind boxyfluorescein succinimidyl ester (CFSE) following a published pro- legs of sacrificed animals were used to harvest cells and prepare tocol.55 CFSE-stained ARI2h cells were then cultured with ARP-1 FACS samples. Mice were assigned to each group randomly, and cells for 3 days at a 0.5:1 ratio of ARI2h:ARP-1 cells and analyzed the experiment was performed in a blinded manner so that both by flow cytometry. Control cultures included ARI2h cells cultured Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 219 Molecular Therapy: Methods & Clinical Development A CD4+ CD8+ PRE POST PRE POST EMRA EMRA Early Early EMRA Early EMRA Early EM EM EM CM* CM* CM EM CM B C + + + + + + + + CD4 CD45RA CCR7 CD8 CD45RA CCR7 CD4 CD8 250 500 15 20 ns ns ns ns 200 400 15 CD25 (MFI x10 ) CD25 (MFI x10 ) 3 3 10 CXCR3 (MFI) CXCR3 (MFI) 150 300 10 100 200 5 5 50 100 0 0 0 0 PRE ARI2 POST ARI2 PRE ARI2 POST ARI2 PRE ARI2 POST ARI2 PRE ARI2 POST ARI2 D CD4 + CD8 + Control (d0) PRE ARI2 POST ARI2 + 10 10 CD4 CD8+ ** * 8 8 CD95 (MFI x10 ) CD95 (MFI x10 ) 3 3 %maximum 6 6 4 4 2 2 0 0 PRE ARI2 POST ARI2 PRE ARI2 POST ARI2 CD95-PE E CD4+ CD8+ PRE ARI2 POST ARI2 PRE ARI2 POST ARI2 300 400 Normalised MFI (% PRE UT) Normalised MFI (% PRE UT) * ns * ns 300 200 ns ns 200 ns ns 100 100 0 0 PD-1 LAG-3 TIM-3 TIGIT PD-1 LAG-3 TIM-3 TIGIT (legend on next page) 220 Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 www.moleculartherapy.org alone, with or without IL-2, and co-cultures that included ARI2h cells stripping buffer (0.2 M glycine [pH 2.5], 0.1% SDS, and 0.1% that were not stained with CFSE. Tween-20) and reprobed with an anti-STAT3 antibody (Cell Signaling Technology). Repeated stimulation of ARI2 h cells with tumor cells We co-cultured 0.125  106 ARI2h cells with 1  106 ARP-1-GFP- Statistical analysis ffLuc cells in a total of 2 mL for 72 h. Thereafter, stimulated ARI2h GraphPad Prism v.8.0.1 (GraphPad Software, San Diego, CA) was cells were combined with 1  106 fresh ARP-1-GFP-ffLuc cells at used for data analysis. For in vitro experiments, comparisons of two the same E:T ratio (0.125:1). ARI2h cells were restimulated with groups were performed using two-tailed paired Student’s t tests, ARP-1-GFP-ffLuc cells three times (a total of four rounds of stimula- and multiple comparisons were conducted using a repeated-measures tion). Following each round of stimulation, ARI2h cells were counted one-way ANOVA corrected with a Tukey post-hoc test. Comparison using an Attune NxT flow cytometer (Thermo Fisher Scientific). of mouse survival between different groups was analyzed using the log Every 24 h, aliquots of the co-cultures were removed to quantify tu- rank (Mantel-Cox) test, and differences in the presence of ARI2h cells mor cell survival, measured as described above. in murine organs was assessed using a two-tailed unpaired Student’s t test. Immunoblotting Monocytes were resuspended in RPMI-1640 media supplemented DATA AVAILABILITY with 10% fetal calf serum (FCS) at 2.5  106/mL and incubated Data are available upon reasonable request. with 1–100 ng/mL recombinant G-CSF or vehicle control for 10 min at 37 C. Cell stimulation was halted by the addition of SUPPLEMENTAL INFORMATION ice-cold PBS and rapid centrifugation. After an additional wash Supplemental information can be found online at https://doi.org/10. in PBS, pellets were frozen and stored at 80 C. Cells were lysed 1016/j.omtm.2022.06.010. in buffer (20 mM Tris [pH 7.4], 150 mM NaCl, 1% Triton X-100, 0.01% SDS, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, ACKNOWLEDGMENTS and 1 mM Tris(2-carboxyethyl)phosphine [TCEP]) supplemented We are extremely grateful to the patients for participating in this with 1% Protease Inhibitor Cocktail, 1% Phosphatase Inhibitor study and to the nurses from the Department of Haematology and Cocktail 2, and 1% Phosphatase Inhibitor Cocktail 3 (all Sigma-Al- the Apheresis Unit at Hospital Clínic, particularly Maria Teresa Sol- drich) for 10 min at 4 C before centrifugation. Clarified lysates ano Moliner, for their invaluable work in obtaining patient samples. were then supplemented with Laemmli Sample Buffer (Bio-Rad) In addition, we acknowledge the Multiple Myeloma Research Center containing 25 mM TCEP and boiled for 5 min. Samples were (Little Rock, AK) for providing the ARP-1 cell line and Amer Najjar resolved by SDS-polyacrylamide gel electrophoresis using 10% for providing the plasmid coding for GFP-ffLuc. Finally, we thank the Mini-PROTEAN TGX Precast Protein Gels (Bio-Rad, Hercules, “la Caixa” Foundation (CP042702), the Spanish Institute of Health CA), and proteins were transferred to polyvinylidene fluoride Carlos III (projects: PI19/00669 and ICI19/00025; co-funded by the (PVDF) membrane. Blots were probed with primary antibodies European Union), and the AECC (LABAE21971FERN) for providing targeting pSTAT3Y705 (Cell Signaling Technology, Danvers, MA) funding for this study. or b-actin (Santa Cruz Biotechnologies, Dallas, TX) before incuba- tion with horseradish-peroxidase-conjugated secondary antibodies AUTHOR CONTRIBUTIONS (goat anti-rabbit [Cell Signaling Technology] or goat anti-mouse Conceptualization, A.M.B., A.U.-I., and C.F.d.L.; data curation, [LI-COR Biosciences, Lincoln, NE]). Membranes were exposed A.M.B.; formal analysis, A.M.B.; funding acquisition, A.U.-I. and to SuperSignal West Pico Chemiluminescent Substrate (Thermo C.F.d.L.; investigation, A.M.B., A.O.-C., M.B.i.C., and D.F.M.; meth- Fisher Scientific), and the resulting chemiluminescence signal odology, A.M.B.; project administration, A.M.B., A.U.-I., and was visualized on an ImageQuant LAS 4000 Mini system (GE C.F.d.L.; resources, M.S.-L., M.L., N.M.-C., J.C., A.U.-I., and Healthcare, Chicago, IL). To detect total STAT3 levels, membranes C.F.d.L.; supervision, C.F.d.L.; visualization, A.M.B.; writing – original initially used to blot for pSTAT3Y705 were stripped with a mild draft, A.M.B.; writing – review & editing, A.M.B., L.G.R.-L., A.U.-I., Figure 8. POST G-CSF ARI2h cells have lower expression of LAG-3 and CD95 (A) Summary of memory and effector phenotypes of CD4+ (left) and CD8+ (right) ARI2h CAR+ T cells, categorized as early differentiated (early; CCR7+CD45RA+), CM (CCR7+- CD45RA), EM (CCR7CD45RA), and effector memory CD45RA+ (CCR7CD45RA+) cells. No populations were significantly different in PRE G-CSF ARI2h and POST G-CSF ARI2h cells, with the exception of the CD4+ CM subset (p = 0.018). Data represent the mean percentage of each population. n = 4. (B) CXCR3 expression on early differentiated (CCR7+CD45RA+) CAR+ CD4+ (left) and CD8+ (right) ARI2h cells is shown, based on the MFI of the staining. (C) Measurement of CD25 expression on CAR+ CD4+ (left) and CD8+ (right) ARI2h cells is shown, displayed as MFI. (D) Relative expression of CD95 on CAR+ CD4+ (left) and CD8+ (right) ARI2h cells is shown. Presented is a summary of the CD95 MFI (left) and representative histograms of the staining (right), which includes unstimulated day 0 T cells as a control. (E) Summary of the surface expres- sion of PD-1, LAG-3, TIM-3, and TIGIT on CAR+ CD4+ (left) and CD8+ (right) ARI2h cells is shown, represented as MFI normalized to that displayed on UT PRE G-CSF cells. Gray shapes indicate data from PRE ARI2h cells, and blue shapes represent data from POST ARI2h cells. Error bars show mean ± SD. Data points in (B)–(E) are coded according to patient donor and are consistent with other figures. **p < 0.01; *p < 0.05. Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 221 Molecular Therapy: Methods & Clinical Development and C.F.d.L. All authors have reviewed and approved the final version multiple myeloma. Blood 128, 384–394. https://doi.org/10.1182/blood-2015-12- 687749. of this manuscript. 16. Chung, D.J., Pronschinske, K.B., Shyer, J.A., Sharma, S., Leung, S., Curran, S.A., Lesokhin, A.M., Devlin, S.M., Giralt, S.A., and Young, J.W. (2016). T-cell exhaustion DECLARATION OF INTERESTS in Multiple myeloma relapse after autotransplant: optimal timing of immunotherapy. The authors declare no competing interests. Cancer Immunol. Res. 4, 61–71. https://doi.org/10.1158/2326-6066.cir-15-0055. 17. Zelle-Rieser, C., Thangavadivel, S., Biedermann, R., Brunner, A., Stoitzner, P., REFERENCES Willenbacher, E., Greil, R., and Jöhrer, K. (2016). T cells in multiple myeloma display 1. Sung, H., Ferlay, J., Siegel, R.L., Laversanne, M., Soerjomataram, I., Jemal, A., and features of exhaustion and senescence at the tumor site. J. Hematol. Oncol. 9, 116. Bray, F. (2021). Global cancer statistics 2020: GLOBOCAN estimates of incidence https://doi.org/10.1186/s13045-016-0345-3. and mortality worldwide for 36 cancers in 185 countries. CA. Cancer J. Clin. 71, 18. Garfall, A.L., Dancy, E.K., Cohen, A.D., Hwang, W.T., Fraietta, J.A., Davis, M.M., 209–249. https://doi.org/10.3322/caac.21660. Levine, B.L., Siegel, D.L., Stadtmauer, E.A., Vogl, D.T., et al. (2019). T-cell phenotypes 2. Shah, U.A., and Mailankody, S. (2020). CAR T and CAR NK cells in multiple associated with effective CAR T-cell therapy in postinduction vs relapsed multiple myeloma: expanding the targets. Best Pract. Res. Clin. Haematol. 33, 101141. myeloma. Blood Adv. 3, 2812–2815. https://doi.org/10.1182/bloodadvances. https://doi.org/10.1016/j.beha.2020.101141. 2019000600. 3. Rodríguez-Lobato, L.G., Ganzetti, M., Fernández de Larrea, C., Hudecek, M., Einsele, 19. Cohen, A.D., Garfall, A.L., Stadtmauer, E.A., Melenhorst, J.J., Lacey, S.F., Lancaster, H., and Danhof, S. (2020). CAR T-cells in multiple myeloma: state of the art and E., Vogl, D.T., Weiss, B.M., Dengel, K., Nelson, A., et al. (2019). B cell maturation future directions. Front. Oncol. 10, 1243. https://doi.org/10.3389/fonc.2020.01243. antigen–specific CAR T cells are clinically active in multiple myeloma. J. Clin. Invest. 129, 2210–2221. https://doi.org/10.1172/jci126397. 4. Bruno, B., Wäsch, R., Engelhardt, M., Gay, F., Giaccone, L., D’Agostino, M., Rodríguez-Lobato, L.G., Danhof, S., Gagelmann, N., Kröger, N., et al. (2021). 20. Bobin, A., Liuu, E., Moya, N., Gruchet, C., Sabirou, F., Lévy, A., Gardeney, H., Nsiala, European Myeloma Network perspective on CAR T-Cell therapies for multiple L., Cailly, L., Guidez, S., et al. (2020). Multiple myeloma: an overview of the current myeloma. Haematologica 106, 2054–2065. https://doi.org/10.3324/haematol.2020. and novel therapeutic approaches in 2020. Cancers 12, 2885. https://doi.org/10.3390/ 276402. cancers12102885. 5. Mullard, A. (2021). FDA approves first BCMA-targeted CAR-T cell therapy. Nat. 21. Bird, S.A., and Boyd, K. (2019). Multiple myeloma: an overview of management. Rev. Drug Discov. 20, 332. https://doi.org/10.1038/d41573-021-00063-1. Palliat. Care Soc. Pract. 13. https://doi.org/10.1177/1178224219868235. 6. Teoh, P.J., and Chng, W.J. (2021). CAR T-cell therapy in multiple myeloma: more 22. Reyes, E., García-Castro, I., Esquivel, F., Hornedo, J., Cortes-Funes, H., Solovera, J., room for improvement. Blood Cancer J. 11, 84. https://doi.org/10.1038/s41408- and Alvarez-Mon, M. (1999). Granulocyte colony-stimulating factor (G-CSF) tran- 021-00469-5. siently suppresses mitogen-stimulated T-cell proliferative response. Br. J. Cancer 7. Berdeja, J.G., Madduri, D., Usmani, S.Z., Jakubowiak, A., Agha, M., Cohen, A.D., 80, 229–235. https://doi.org/10.1038/sj.bjc.6690344. Stewart, A.K., Hari, P., Htut, M., Lesokhin, A., et al. (2021). Ciltacabtagene autoleucel, 23. Bunse, C.E., Borchers, S., Varanasi, P.R., Tischer, S., Figueiredo, C., Immenschuh, S., a B-cell maturation antigen-directed chimeric antigen receptor T-cell therapy in pa- Kalinke, U., Köhl, U., Goudeva, L., Maecker-Kolhoff, B., et al. (2013). Impaired func- tients with relapsed or refractory multiple myeloma (CARTITUDE-1): a phase 1b/2 tionality of antiviral T cells in G-CSF mobilized stem cell donors: implications for the open-label study. Lancet 398, 314–324. https://doi.org/10.1016/s0140-6736(21) selection of CTL donor. PLoS One 8, e77925. https://doi.org/10.1371/journal.pone. 00933-8. 0077925. 8. Ackley, J., Ochoa, M.A., Ghoshal, D., Roy, K., Lonial, S., and Boise, L.H. (2021). 24. Bunse, C.E., Tischer, S., Lahrberg, J., Oelke, M., Figueiredo, C., Blasczyk, R., and Eiz- Keeping myeloma in check: the past, present and future of immunotherapy in mul- Vesper, B. (2016). Granulocyte colony-stimulating factor impairs CD8+ T cell func- tiple myeloma. Cancers 13, 4787. https://doi.org/10.3390/cancers13194787. tionality by interfering with central activation elements. Clin. Exp. Immunol. 185, 9. Munshi, N.C., Anderson, L.D., Shah, N., Madduri, D., Berdeja, J., Lonial, S., Raje, N., 107–118. https://doi.org/10.1111/cei.12794. Lin, Y., Siegel, D., Oriol, A., et al. (2021). Idecabtagene vicleucel in relapsed and re- 25. Franzke, A., Piao, W., Lauber, J., Gatzlaff, P., Könecke, C., Hansen, W., Schmitt- fractory multiple myeloma. N. Engl. J. Med. 384, 705–716. https://doi.org/10.1056/ Thomsen, A., Hertenstein, B., Buer, J., and Ganser, A. (2003). G-CSF as immune nejmoa2024850. regulator in T cells expressing the G-CSF receptor: implications for transplantation 10. Pratt, G., Goodyear, O., and Moss, P. (2007). Immunodeficiency and immunotherapy and autoimmune diseases. Blood 102, 734–739. https://doi.org/10.1182/blood- in multiple myeloma. Br. J. Haematol. 138, 563–579. https://doi.org/10.1111/j.1365- 2002-04-1200. 2141.2007.06705.x. 26. Pan, L., Delmonte, J., Jalonen, C.K., and Ferrara, J.L.M. (1995). Pretreatment of donor 11. Zavidij, O., Haradhvala, N.J., Mouhieddine, T.H., Sklavenitis-Pistofidis, R., Cai, S., mice with granulocyte colony-stimulating factor polarizes donor T lymphocytes to- Reidy, M., Rahmat, M., Flaifel, A., Ferland, B., Su, N.K., et al. (2020). Single-cell ward type-2 cytokine production and reduces severity of experimental graft- RNA sequencing reveals compromised immune microenvironment in precursor versus-host disease. Blood 86, 4422–4429. https://doi.org/10.1182/blood.v86.12. stages of multiple myeloma. Nat. Cancer 1, 493–506. https://doi.org/10.1038/ 4422.bloodjournal86124422. s43018-020-0053-3. 27. Sloand, E.M., Kim, S., Maciejewski, J.P., Van Rhee, F., Chaudhuri, A., Barrett, J., and 12. Zanwar, S., Nandakumar, B., and Kumar, S. (2020). Immune-based therapies in the Young, N.S. (2000). Pharmacologic doses of granulocyte colony-stimulating factor management of multiple myeloma. Blood Cancer J. 10, 84. https://doi.org/10.1038/ affect cytokine production by lymphocytes in vitro and in vivo. Blood 95, 2269– s41408-020-00350-x. 2274. https://doi.org/10.1182/blood.v95.7.2269. 13. Giles, A.J., Hutchinson, M.K.N.D., Sonnemann, H.M., Jung, J., Fecci, P.E., Ratnam, 28. Morris, E.S., MacDonald, K.P.A., Rowe, V., Johnson, D.H., Banovic, T., Clouston, N.M., Zhang, W., Song, H., Bailey, R., Davis, D., et al. (2018). Dexamethasone- A.D., and Hill, G.R. (2004). Donor treatment with pegylated G-CSF augments the induced immunosuppression: mechanisms and implications for immunotherapy. generation of IL-10-producing regulatory T cells and promotes transplantation toler- J. Immunother. Cancer 6, 51. https://doi.org/10.1186/s40425-018-0371-5. ance. Blood 103, 3573–3581. https://doi.org/10.1182/blood-2003-08-2864. 14. Cooke, R.E., Quinn, K.M., Quach, H., Harrison, S., Prince, H.M., Koldej, R., and 29. Künkele, A., Brown, C., Beebe, A., Mgebroff, S., Johnson, A.J., Taraseviciute, A., Ritchie, D. (2020). Conventional treatment for multiple myeloma drives premature Rolczynski, L.S., Chang, C.A., Finney, O.C., Park, J.R., and Jensen, M.C. (2019). aging phenotypes and metabolic dysfunction in T cells. Front. Immunol. 11, 2153. Manufacture of chimeric antigen receptor T cells from mobilized cyropreserved pe- https://doi.org/10.3389/fimmu.2020.02153. ripheral blood stem cell units depends on monocyte depletion. Biol. Blood Marrow Transplant. 25, 223–232. https://doi.org/10.1016/j.bbmt.2018.10.004. 15. Krejcik, J., Casneuf, T., Nijhof, I.S., Verbist, B., Bald, J., Plesner, T., Syed, K., Liu, K., Van De Donk, N.W.C.J., Weiss, B.M., et al. (2016). Daratumumab depletes CD38+ 30. Wang, X., Borquez-Ojeda, O., Stefanski, J., Du, F., Qu, J., Chaudhari, J., Thummar, K., immune regulatory cells, promotes T-cell expansion, and skews T-cell repertoire in Zhu, M., Shen, L.B., Hall, M., et al. (2021). Depletion of high-content CD14+ cells 222 Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 www.moleculartherapy.org from apheresis products is critical for successful transduction and expansion of CAR mechanism for monocyte-mediated suppression of T- cell proliferation. Blood 92, T cells during large-scale cGMP manufacturing. Mol. Ther. Methods Clin. Dev. 22, 215–222. https://doi.org/10.1182/blood.v92.1.215.413k10_215_222. 377–387. https://doi.org/10.1016/j.omtm.2021.06.014. 44. Wang, K., Lv, M., Chang, Y.J., Zhao, X.Y., Zhao, X.S., Zhang, Y.Y., Sun, Y.Q., Wang, 31. Perez-Amill, L., Suñe, G., Antoñana-Vildosola, A., Castella, M., Najjar, A., Bonet, J., Z.D., Suo, P., Zhou, Y., et al. (2019). Early myeloid-derived suppressor cells (HLA- Fernández-Fuentes, N., Inogés, S., López, A., Bueno, C., et al. (2021). Preclinical DR-/lowCD33+CD16-) expanded by granulocyte colony-stimulating factor prevent development of a humanized chimeric antigen receptor against B cell maturation an- acute graft-versus-host disease (GVHD) in humanized mouse and might contribute tigen for multiple myeloma. Haematologica 106, 173–184. https://doi.org/10.3324/ to lower GVHD in patients post allo-HSCT. J. Hematol. Oncol. 12, 31. https://doi.org/ haematol.2019.228577. 10.1186/s13045-019-0710-0. 32. Oliver-Caldes, A., Jiménez, R., Español-Rego, M., Cibeira, M.T., Ortiz-Maldonado, 45. Toh, H.C., Sun, L., Soe, Y., Wu, Y., Phoon, Y.P., Chia, W.K., Wu, J., Wong, K.Y., and V., Quintana, L.F., Castillo, P., Guijarro, F., Tovar, N., Montoro, M., et al. (2021). Tan, P. (2009). G-CSF induces a potentially tolerant gene and immunophenotype First report of CART treatment in AL amyloidosis and relapsed/refractory multiple profile in T cells in vivo. Clin. Immunol. 132, 83–92. https://doi.org/10.1016/j.clim. myeloma. J. Immunother. Cancer 9, e003783. https://doi.org/10.1136/jitc-2021- 2009.03.509. 003783. 46. Castella, M., Boronat, A., Martín-Ibáñez, R., Rodríguez, V., Suñé, G., Caballero, M., 33. Theron, A.J., Steel, H.C., Rapoport, B.L., and Anderson, R. (2020). Contrasting im- Marzal, B., Pérez-Amill, L., Martín-Antonio, B., Castaño, J., et al. (2019). munopathogenic and therapeutic roles of granulocyte colony-stimulating factor in Development of a novel anti-CD19 chimeric antigen receptor: a paradigm for an cancer. Pharmaceuticals 13, 406. https://doi.org/10.3390/ph13110406. affordable CAR T cell production at academic institutions. Mol. Ther. Methods 34. Nishiki, S., Hato, F., Kamata, N., Sakamoto, E., Hasegawa, T., Kimura-Eto, A., Hino, Clin. Dev. 12, 134–144. https://doi.org/10.1016/j.omtm.2018.11.010. M., and Kitagawa, S. (2004). Selective activation of STAT3 in human monocytes stim- ulated by G-CSF: implication in inhibition of LPS-induced TNF-a production. Am. J. 47. Gattinoni, L., Speiser, D.E., Lichterfeld, M., and Bonini, C. (2017). T memory stem Physiol. Cell Physiol. 286, 1302–1311. https://doi.org/10.1152/ajpcell.00387.2003. cells in health and disease. Nat. Med. 23, 18–27. https://doi.org/10.1038/nm.4241. 35. Frigault, M.J., Lee, J., Basil, M.C., Carpenito, C., Motohashi, S., Scholler, J., Kawalekar, 48. Costello, C.L., Cohen, A.D., Patel, K.K., Ali, S.S., Berdeja, J.G., Shah, N., Ganguly, S., O.U., Guedan, S., McGettigan, S.E., Posey, A.D., et al. (2015). Identification of Kocoglu, M.H., Abedi, M., Ostertag, E.M., et al. (2020). Phase 1/2 study of the safety chimeric antigen receptors that mediate constitutive or inducible proliferation of and response of P-BCMA-101 CAR-T cells in patients with relapsed/refractory (r/r) T cells. Cancer Immunol. Res. 3, 356–367. https://doi.org/10.1158/2326-6066.cir- multiple myeloma (MM) (PRIME) with novel therapeutic strategies. Blood 136, 14-0186. 29–30. https://doi.org/10.1182/blood-2020-142695. 36. Battram, A.M., Bachiller, M., and Martín-Antonio, B. (2020). Senescence in the devel- 49. Turtle, C.J., Hanafi, L.A., Berger, C., Gooley, T.A., Cherian, S., Hudecek, M., opment and response to cancer with immunotherapy: a double-edged sword. Int. J. Sommermeyer, D., Melville, K., Pender, B., Budiarto, T.M., et al. (2016). CD19 Mol. Sci. 21, 4346. https://doi.org/10.3390/ijms21124346. CAR-T cells of defined CD4+:CD8+ composition in adult B cell ALL patients. 37. Gattinoni, L., Lugli, E., Ji, Y., Pos, Z., Paulos, C.M., Quigley, M.F., Almeida, J.R., J. Clin. Invest. 126, 2123–2138. https://doi.org/10.1172/jci85309. Gostick, E., Yu, Z., Carpenito, C., et al. (2011). A human memory T cell subset 50. Sommermeyer, D., Hudecek, M., Kosasih, P.L., Gogishvili, T., Maloney, D.G., Turtle, with stem cell-like properties. Nat. Med. 17, 1290–1297. https://doi.org/10.1038/ C.J., and Riddell, S.R. (2016). Chimeric antigen receptor-modified T cells derived nm.2446. from defined CD8+ and CD4+ subsets confer superior antitumor reactivity in vivo. 38. Fraietta, J.A., Lacey, S.F., Orlando, E.J., Pruteanu-Malinici, I., Gohil, M., Lundh, S., Leukemia 30, 492–500. https://doi.org/10.1038/leu.2015.247. Boesteanu, A.C., Wang, Y., O’Connor, R.S., Hwang, W.T., et al. (2018). 51. Deng, Q., Han, G., Puebla-Osorio, N., Ma, M.C.J., Strati, P., Chasen, B., Dai, E., Dang, Determinants of response and resistance to CD19 chimeric antigen receptor (CAR) M., Jain, N., Yang, H., et al. (2020). Characteristics of anti-CD19 CAR T cell infusion T cell therapy of chronic lymphocytic leukemia. Nat. Med. 24, 563–571. https:// products associated with efficacy and toxicity in patients with large B cell lymphomas. doi.org/10.1038/s41591-018-0010-1. Nat. Med. 26, 1878–1887. https://doi.org/10.1038/s41591-020-1061-7. 39. Lécuroux, C., Girault, I., Urrutia, A., Doisne, J.M., Deveau, C., Goujard, C., Meyer, L., Sinet, M., and Venet, A. (2009). Identification of a particular HIV-specific CD8+ 52. Battram, A.M., Bachiller, M., Lopez, V., Fernández de Larrea, C., Urbano-Ispizua, A., T-cell subset with a CD27+ CD45RO/RA+ phenotype and memory characteristics af- and Martín-Antonio, B. (2021). IL-15 enhances the persistence and function of ter initiation of HAART during acute primary HIV infection. Blood 113, 3209–3217. BCMA-targeting CAR-T cells compared to IL-2 or IL-15/IL-7 by limiting CAR-T https://doi.org/10.1182/blood-2008-07-167601. cell dysfunction and differentiation. Cancers 13, 3534. https://doi.org/10.3390/can- cers13143534. 40. Arpinati, M., Green, C.L., Heimfeld, S., Heuser, J.E., and Anasetti, C. (2000). Granulocyte-colony stimulating factor mobilizes T helper 2-inducing dendritic cells. 53. Bachiller, M., Perez-Amill, L., Battram, A.M., Carné, S.C., Najjar, A., Verhoeyen, E., Blood 95, 2484–2490. https://doi.org/10.1182/blood.v95.8.2484. Juan, M., Urbano-Ispizua, A., and Martin-Antonio, B. (2021). NK cells enhance 41. Mielcarek, M., Martin, P.J., and Torok-Storb, B. (1997). Suppression of alloantigen- CAR-T cell antitumor efficacy by enhancing immune/tumor cells cluster formation induced T-cell proliferation by CD14+ cells derived from granulocyte colony-stimu- and improving CAR-T cell fitness. J. Immunother. Cancer 9, e002866. https://doi. lating factor-mobilized peripheral blood mononuclear cells. Blood 89, 1629–1634. org/10.1136/jitc-2021-002866. https://doi.org/10.1182/blood.v89.5.1629.1629_1629_1634. 54. Lucas, F., Pennell, M., Huang, Y., Benson, D.M., Efebera, Y.A., Chaudhry, M., 42. Luyckx, A., Schouppe, E., Rutgeerts, O., Lenaerts, C., Fevery, S., Devos, T., Dierickx, Hughes, T., Woyach, J.A., Byrd, J.C., Zhang, S., et al. (2020). T cell transcriptional D., Waer, M., Van Ginderachter, J.A., and Billiau, A.D. (2012). G-CSF stem cell mobi- profiling and immunophenotyping uncover LAG3 as a potential significant target lization in human donors induces polymorphonuclear and mononuclear myeloid- of immune modulation in multiple myeloma. Biol. Blood Marrow Transplant. 26, derived suppressor cells. Clin. Immunol. 143, 83–87. https://doi.org/10.1016/j.clim. 7–15. https://doi.org/10.1016/j.bbmt.2019.08.009. 2012.01.011. 55. Marchingo, J.M., Sinclair, L.V., Howden, A.J., and Cantrell, D.A. (2020). Quantitative 43. Mielcarek, M., Graf, L., Johnson, G., and Torok-Storb, B. (1998). Production of inter- analysis of how myc controls t cell proteomes and metabolic pathways during t cell leukin-10 by granulocyte colony-stimulating factor- mobilized blood products: a activation. Elife 9, e53725. https://doi.org/10.7554/elife.53725. Molecular Therapy: Methods & Clinical Development Vol. 26 September 2022 223 OMTM, Volume 26 Supplemental information T cells isolated from G-CSF-treated multiple myeloma patients are suitable for the generation of BCMA-directed CAR-T cells Anthony M. Battram, Aina Oliver-Caldés, Maria Suárez-Lledó, Miquel Lozano, Miquel Bosch i Crespo, Núria Martínez-Cibrián, Joan Cid, David F. Moreno, Luis Gerardo Rodríguez-Lobato, Alvaro Urbano-Ispizua, and Carlos Fernández de Larrea Table S1. Characteristics of multiple myeloma patients (n = 8). Parameter Value, n Median age (range), years 62 (44-66) Sex Male 1 Female 7 ISSa I 1 II 5 III 2 R-ISSb I 1 II 2 III 5 Isotype IgG 5 IgA 1 Light chain only 2 Cytogeneticsc Del17p 3 t(11;14) 1 IGH rearrangement 2 Hyperdiploidy 1 No adverse cytogenetics 3 Median bone marrow plasma 53 (16-85) cell infiltration (range), % Extramedullary disease Yes 0 No 8 Median creatinine at diagnosis 0.74 (0.56-1.01) (range), mg/dL Lines of previous therapy 1 8 Stem cell mobilisation (days) 4 8 a ISS, International Staging System b R-ISS, Revised International Staging System c n>8 as some patients had multiple adverse cytogenetic characteristics 1 Table S2. Induction therapy (n = 8). Parameter Value, n a Induction therapy regimen VRD 2 KRD 6 b Response to therapy VGPR 4 CR 4 Cycles of therapy 4 5 5 3 a VRD, bortezomib/lenalidomide/dexamethasone; KRD, carfilzomib/lenalidomide/ dexamethasone b VGPR, very good partial response; CR, complete response 2 Figure S1. Recombinant G-CSF induces STAT3 phosphorylation in monocytes Monocytes treated with 0-100 ng/ml recombinant G-CSF for 10 minutes were subject to Western blotting to assess STAT3 phosphorylation. Membranes were reprobed for total STAT3 and β-actin was used as a loading control. Numbers on the left of the blots show molecular weights in kDa. Results are representative of 2 independent experiments. Figure S2. Staining of T cells to assess apoptosis Cells were stained with annexin V (AnV) and 7-AAD on day -3 before exposure to G-CSF and afterwards on day 0 to determine the frequency of live cells (AnV-7-AAD-), early apoptotic cells (AnV+7-AAD-) and late apoptotic/dead cells (AnV+7-AAD+). Numbers indicate frequency of each population compared to total cells. See also Figure 1A-B. 3 Figure S3. T-cell isolation from PBMCs successfully removes monocytes T cells freshly isolated from patients before (PRE) and after (POST) G-CSF administration were analysed for the presence of T cells and monocytes by staining for CD3 and CD14, respectively. PBMCs from POST G-CSF samples before T-cell isolation were also stained as a control. Numbers indicate frequency of each population compared to total cells. 4 Figure S4. CD8+ ARI2h cell granzyme B production and proliferation of CD4+ ARI2h cells in response to myeloma cells are unaffected by G-CSF exposure (A) UT and ARI2h cells were co-incubated with ARP-1 cells or media alone (alone) for 6 hours. Expression of granzyme B in CD8+ CAR+ cells was assessed by flow cytometry and displayed are histograms representative of 3 experiments. FMO indicates fluorescence minus one (fully stained except for granzyme B) control and numbers on histograms show MFI. See also Figure 6D. (B) ARI2h cells were labelled with CFSE and cultured with (+ARP-1) or without (alone) ARP-1 cells. After 72 hours, ARI2h cell proliferation was analysed by measuring CFSE dilution. ARI2h cells that were not treated with CFSE (Unlabelled) acted as a control. The displayed histograms are CD4+ CAR+ cells and are representative of 3 experiments. See also Figure 6E. 5 Figure S5. PRE G-CSF ARI2h and POST G-CSF ARI2h cells respond similarly to repeated stimulation with tumour cells 0.125x106 ARI2h cells were mixed with 1x106 GFP-ffLuc-expressing ARP-1 (ARP-1-GFP) cells. 72 hours later (day +3), ARI2h cells were counted and combined with fresh ARP-1-GFP cells at the same ARI2h cell:ARP-1- GFP cell ratio (0.125:1). This tumour cell stimulation was performed a total of four times. See also Figure 6F. (A) Schematic of in vitro repeated stimulation experimental design. (B) Survival of ARP-1-GFP cells during the first (Challenge 1), second (Challenge 2) and third (Challenge 3) stimulations, determined by measuring target cell bioluminescence 6 (n = 3). Data are represented as mean ± SEM. See also Figure 6G. (C-F) Percentage of T cells that were positive for the ARI2h CAR molecule (C), CCR7/CD45RA staining of CD4+ and CD8+ CAR+ cells (D), CD27/CD45RA staining of CD8+ CAR+ cells (E) and summary of exhaustion marker expression (F) after all four stimulations. FACS plots in panels D and E are representative of 3 experiments and data points in panel F are coded according to patient donor and are consistent with other figures. Error bars show mean ± SD. Figure S6. PRE and POST G-CSF ARI2h cells are similarly effective in a mouse MM model In vivo models were performed as shown in Figure 7A. Tumour progression was monitored by weekly readings of animal bioluminescence. Shown are photos taken from the dorsal side of the mice. Numbers on the left of the image show the number of days following tumour cell infusion. See also Figure 7B. 7 Figure S7. LAG-3 expression is reduced on POST G-CSF ARI2h cells Relative expression of LAG-3 on CAR+ CD4+ (left) and CD8+ (right) ARI2h cells and unstimulated day 0 T cells as a control. Histograms are representative of 5 experiments. See also Figure 8E. 8 Figure S8. Gating strategy to quantify CD4+CD25highFoxP3+ Tregs CD4+ conventional T cells (CD25lowCD127high) were used to generate a gate to identify CD25highFoxP3+ Tregs from the total population of CD4+ T cells. 9