Cell composition and expansion strategy can reduce the beneficial effect of AKT-inhibition on functionality of CD8⁺ T cells

PBMCs were isolated from buffy coats of healthy donors (Sanquin Blood Supply Foundation) using Ficoll-paque™ PLUS (cat#17-1440-03, Sigma-Aldrich) density gradient centrifugation. CD14⁺ monocytes were isolated from fresh material using anti-CD14 immunomagnetic beads (cat#130-050-201, Miltenyi Biotec). Peripheral blood lymphocytes were cryopreserved in medium containing 10% (cat#102,952, DMSO, Merck Millipore) and 10% serum using a Mr.Frosty™ (Thermofisher). T cells were isolated from fresh or frozen material via magnetic bead isolation using the EasySep™ Human Naive CD4⁺ T Cell Isolation Kit (cat#19555, Stemcell Technologies) or CD3, CD4 or CD8 microbeads (cat#130-050-101, cat#130-045-101, and cat#130-045-201, Miltenyi Biotec). The obtained purity was on average 95% (ranging from 91 to 99%) and CD3, CD4 and CD8 bead-isolated T cells contained on average 52% T_N cells (ranging from 27 to 83%). Cell numbers were based on trypan blue counting and flow cytometric quantification of relative and absolute cell numbers.

To generate mature DCs, monocytes were cultured for 2–3 days in manufactory pre-tested X-VIVO™ 15 medium (cat#BE02-061Q or cat#BE02-060Q, Lonza) supplemented with 2% human serum (HS, Sanquin Blood Supply Foundation, not pre-tested for assay performance), 500 IU/ml IL4 and 800 IU/ml GM-CSF (cat#11340045 and cat#11343125, ImmunoTools). Subsequently, cells were re-plated at 0.5 × 10⁶/ml in fresh X-VIVO™ 15 containing 2% HS, 500 IU/ml IL4 and 800 IU/ml GM-CSF for another 3–4 days. Next, immature DCs were maturated for 2 days by adding X-VIVO™ 15 containing 2% HS and cytokines at a final concentration of 5 ng/ml IL1β, 15 ng/ml IL6, 20 ng/ml TNFα (cat#11340015, cat#11340064, cat#11343015, all ImmunoTools) and 1 µg/ml PGE2 (Prostin E2®, Pfizer).

CD3⁺ T cells, CD4⁺ T cells, CD4⁺ T_N or CD8⁺ T cells were stimulated with allogeneic DCs at a 1:10 DC:T cell ratio or with CD3/CD28 Dynabeads (cat#11131D, ThermoFisher) at a 1:1 bead: T cell ratio for 10 days in IMDM (cat#12440053, Thermofisher), supplemented with 10% HS, 50 IU/ml IL2 (Proleukin®, Chiron), 5 ng/ml IL7 and 5 ng/ml IL15 (cat#11340075 and cat#11340155, both ImmunoTools). When indicated, Akt-inhibitor VIII (cat#A6730, Sigma) or GDC-0068 (cat#HY-15186, MedChemExpress), dissolved in DMSO, was added. Control conditions were supplemented with corresponding concentrations of DMSO (≤ 0.1% in all experiments). Half of the culture volume was refreshed with medium plus cytokines and AKT-inhibitor or DMSO every 2–3 days. Cells were rechallenged with DCs or CD3/CD28 beads on day 10 and cultured for another 7 days without AKT-inhibitor or DMSO. To evaluate cytokine production profiles on day 10 or 17, T cells were restimulated overnight with DCs (1:10 DC:T cell ratio) or 1 µg/ml PMA plus 40 µg/ml ionomycin (cat# P1585 and cat#I0634, both Sigma), in the presence of Brefeldin A (Golgiplug, cat#555029, BD Biosciences) or Monensin (Golgistop, cat#554724 BD Biosciences), in absence of AKT-inhibitor or DMSO.

Immunophenotypical analysis was performed by flow cytometry. Cells were washed with PBS (cat#3623130, Braun) containing 0.5% bovine serum albumin (BSA, cat#A9647, Sigma) followed by extracellular staining for 30 min at 4 °C in the dark with antibodies against CD3 (UCHT1, Biolegend), CD4 (SK3, Biolegend), CD8 (RPA-T8, Biolegend), CD45RO (UCHL1, Beckman Coulter) and Fixable Viability Dye eFluor780 (cat#65-0865-14, ThermoFisher). Additionally, for phenotyping the following antibodies were used: CCR7 (G04H7, Biolegend), CD62L (DREG56, Biolegend), CXCR4 (12G5, Biolegend), CXCR3 (G025H7, Biolegend), CCR4 (L291H4, Biolegend), CCR6 (G034E3, Biolegend) and CD25 (B1.49.9, Beckman Coulter). For intracellular transcription factor staining, cells were fixated and permeabilized with Fixation/Permeabilization Concentrate and Diluent (cat#00-5123-43, and cat#005223-56, both eBioscience) and stained in Permeabilization Buffer (cat#00-8333-56, eBioscience) with antibodies for T-bet (4B10, eBioscience), GATA3 (16E10A23, Biolegend), RORγT (AFKJS-9, eBioscience) and FOXP3 (3G3, eBioscience). Intracellular staining of CD4⁺ T cell cytokines was performed upon fixation and permeabilization with Fixation/Permeabilization Concentrate and Diluent after overnight restimulation in presence of Brefeldin A to analyze IFNγ (B27, BD Biosciences), IL4 (8D4-8, BD Biosciences) and IL17 (BL168, Biolegend), or Monensin to analyze IL10 (JES3-9D7, BD Biosciences). CD8⁺ T cell polyfunctionality was analyzed after overnight stimulation in presence of Brefeldin A and CD107a (H4A3, Biolegend) after which cells were fixated with 4% paraformaldehyde (cat#P6148, BOOM) for 10 min at RT in the dark, followed by permeabilization with 0.1% saponin (cat#47036, Sigma) buffer containing 10% FCS (Integro B.V.) for 10 min at RT. Staining of IFNγ, IL2 (5344.111, BD Biosciences) and TNFα (MAb11, BD Biosciences) was performed in 0.1% saponin buffer for 30 min at 4 °C in the dark. After staining, cell data acquisition was performed on the Gallios flow cytometer (Beckman Coulter). Flow cytometry data were analyzed with Kaluza software (Beckman Coulter, version 1.5a and 2.1). Delta median fluorescence intensity (ΔMFI) was calculated by subtracting the MFI of the marker of interest with the MFI of the background fluorescence within the population of interest. Polyfunctionality of CD8⁺ T cells was visualized using SPICE version 5.3 (NIH/NIAID, MD, USA).

This paper is MIATA-compliant. This study was conducted in a laboratory that operates under exploratory research principles. Experiments are performed in general research investigative assays according to investigative protocols. Raw data can be provided per request. The MIATA checklist is provided as Supplementary information.

We previously showed that T_SCM-like CD8⁺ T cells can be generated by transient AKT-inhibition during ex vivo T cell activation and expansion by allo-antigen or minor histocompatibility antigen-presenting DCs [7, 11]. To study the effect of AKT-inhibition on effector memory differentiation of CD4⁺ T cells, we stimulated CD4 T_N cells with allogeneic DCs in combination with increasing dosages of AKT-inhibitor. First, DMSO drug solvent effects were excluded, as viability, activation, expansion and differentiation of CD4⁺ T cells were not affected (Supplementary Fig. 1). Notably, increasing dosages of AKT-inhibitor did not affect viability (Fig. 1a) nor activation (Fig. 1b), though a dose-dependent effect on expansion was observed (Fig. 1c). CD4⁺ T cell expansion reduced from 40-fold in control conditions to 20-fold at the highest dose of AktiVIII (18 µM) and GDC (10 µM) during the 10-day culture period. Importantly, the naive-associated marker CD62L was dose-dependently retained on CD4⁺ T cells cultured with AKT-inhibitors (Fig. 1d, e). In addition, CCR7 and CXCR4 were preserved on AktiVIII-treated CD4⁺ T cells, while these effects were less prominent for GDC-treated CD4⁺ T cells. Especially, CCR7 could not be preserved by this ATP-competitive AKT-inhibitor. Overall, ex vivo AKT-inhibition preserved memory differentiation of CD4⁺ T_N cells upon stimulation in a dose-dependent manner, with the strongest effects shown for the allosteric inhibitor AktiVIII.

Besides effector memory differentiation, CD4⁺ T cell develop toward T helper subsets with distinct effector functions upon activation of naive T cells. Therefore, the effect of AKT-inhibition on subset skewing toward Th1, Th2, Th17 and Tregs was evaluated based on extracellular phenotype, intracellular expression of transcription factors and cytokine secretion profiles. The Th17-associated markers (CCR6, RORγT and IL17) were not found upon activation of CD4⁺ T_N cells and therefore not further evaluated in this assay. Remarkably, while the Th1 population was most dominant in control cultures, extracellular expression of the Th1-associated marker CXCR3 was lower on AKT-inhibited CD4⁺ T cells (Fig. 2a, d; Supplementary Fig. 2A). On the contrary, Th2-associated CCR4 was significantly higher expressed on AktiVIII-cultured cells (Fig. 2a, d and Supplementary Fig. 2A). Effects on transcription factor expression were less pronounced, with only significant higher levels of Th2-associated GATA3 in AktiVIII cultured cells (Fig. 2b, d and Supplementary Fig. 2B). Additionally, cytokine secretion profiles were analyzed after overnight restimulation with allogeneic DCs. AKT-inhibited cultures showed significantly lower percentages of IFNγ-producing T cells, confirming the inhibitory effect of AKT-inhibition on Th1 differentiation (Fig. 2c, d and Supplementary Fig. 2C). Again, skewing toward the Th2 subset was observed as reflected by the increased percentage of AKT-inhibited T cells producing IL4. Although the percentage of IL10-producing cells was low (≤ 3%), a small increase in this Treg-associated cytokine was observed for GDC-cultured conditions. Together, these data indicate that both AKT-inhibitors AktiVIII and GDC markedly skew the acquisition of CD4⁺ effector functions toward Th2 traits and less toward Th1 cell differentiation. Moreover, the total proportion of cytokine producing CD4⁺ T cells was significantly lower in AKT-inhibitor-treated cultures (Fig. 2e). Notably, when cells were released from their transient AKT-inhibition and cultured for an additional week, the negative effect on Th1 skewing was sustained, as IFNγ-production remained significantly lower in AktiVIII-treated cultures (Fig. 2f). Together, these data show that in vitro AKT-inhibition promotes CD4⁺ Th-subset skewing toward Th2-associated cells while diminishing Th1 skewing.

As manufacturing of most adoptive T cell products starts with total CD3⁺ T cells, containing total CD4⁺ instead of CD4⁺ T_N cells, we investigated the effect of AKT-inhibition on CD4⁺ T cell differentiation when starting with either total CD3⁺ or CD4⁺ T cell cultures. Similar to starting from CD4⁺ T_N cells, preserved memory differentiation in AKT-inhibited CD4⁺ T cells was observed (Fig. 3a and Supplementary Fig. 3). The cell composition of CD3⁺ T cell cultures consisted of a CD4:CD8 ratio of 2.0 ± 0.5, which shifted to a more balanced culture with a ratio of 0.9 ± 0.2 on day 10. The addition of AKT-inhibition did not affect this distribution. Interestingly, when both CD4⁺ and CD8⁺ T cells were present in the culture (started from CD3⁺ T cells), fewer cytokine-secreting CD4⁺ T cells were observed compared to cultures started from CD4⁺ T cells alone (Fig. 3b–d). Here, 38% of CD4⁺ T cells in single cultures were capable of secreting either IFNγ, IL4, TNFα and/or IL10 while in CD3⁺ T cells-derived cultures only 11% cytokine-secreting CD4⁺ T cells were observed. When AKT-inhibition was added to these cultures, secretion of the Th-associated cytokines was similar between inhibited cultures, independent of the cell composition. Compared to their controls, AKT-inhibition impeded the proportion of CD4⁺ T cells producing IFNγ when cultures were started with CD4⁺ T cells, though an increase was observed compared to controls of CD3⁺ T cell cultures (Fig. 3b, c). Moreover, also negative effects on CXCR3 and T-bet expression were less pronounced and more variable in total CD3⁺ T cell cultures compared to cultures started with solely CD4⁺ T cells (Supplementary Fig. 4). Nevertheless, independent of the cell composition, skewing toward Th2 cells was observed upon AKT-inhibition (Fig. 3b, c and Supplementary Fig. 4). Together, AKT-inhibition resulted in a decreased Th1-associated IFNγ/Th2-associated IL4 ratio (CD4⁺ T cell cultures: from 6.6 ± 1.5 in control conditions to 2.1 ± 0.5 and 2.9 ± 0.5 in AktiVIII and GDC-treated cultures, respectively, both p < 0.001; CD3⁺ T cell cultures: from 3.2 ± 0.6 in control conditions to 1.7 ± 0.5 and 2.3 ± 0.3 in AktiVIII and GDC-treated cultures, respectively). Moreover, IL17-producing CD4⁺ T cells could be detected when starting cultures with either total CD3⁺ or CD4⁺ T cells, with increased IL17-producing CD4⁺ T cells upon AKT-inhibition compared to control CD3⁺ T cells (Fig. 3b, c). Lastly, IL10-production was again low (≤ 3%) in these cultures, with no significant effect on IL10-producing cells upon AKT-inhibition. Furthermore, cell composition did not affect the expansion capacity of the cultured cells (Fig. 3e). Nevertheless, upon antigen recall in the absence of inhibitors a mild increase in expansion capacity of AKT-inhibited CD4⁺ T cells as was observed. Collectively, these data show that cell composition determine CD4⁺ T helper skewing and the effect of AKT-inhibition.

As changes in the CD4⁺ Th-skewing could affect CD8⁺ T cell functionality, which is pivotal for effective anti-tumor efficacy, we analyzed the effect of AKT-inhibition on CD8⁺ T cell functionality in presence and absence of CD4⁺ T cells. First, we confirmed the preserved early memory phenotype of CD8⁺ T cells in AktiVIII and GDC-cultured conditions when starting with total CD3⁺ or CD8⁺ T cells based on CD62L expression (Fig. 4a). Notably, in (non-AKT-inhibited) control conditions the presence of CD4⁺ T cells reduced CD8⁺ T cell differentiation, shown by higher CD62L expression compared to cultures with CD8⁺ T cells alone. Furthermore, in control conditions the presence of CD4⁺ T cells also improved the rechallenge expansion capacity of control CD8⁺ T cells upon antigen recall as was expected (Fig. 4b). Importantly, transient AKT-inhibition provided significantly enhanced CD8⁺ T cell expansion upon rechallenge, though effects were less pronounced when CD4⁺ T cells were present in the starting material. To that end, we observed preserved memory differentiation of AKT-inhibited CD8⁺ T cells, though cell composition did affect memory phenotype and rechallenge expansion capacity.

As killing capacity and cytokine production of CD8⁺ T cells is pivotal for their function in anti-tumor immunity, we further analyzed CD8⁺ T cell functionality on degranulation and polyfunctional effector cytokine production after one week of rechallenge in absence of AKT-inhibitors. AKT-inhibited CD8⁺ T cells showed an improved degranulation compared to controls, but only when the expansion was performed in the absence of CD4⁺ T cells (Fig. 5a). Moreover, we observed an enhanced production of IFNγ, IL2 and TNFα by allogeneic-DC expanded AKT-inhibited CD8⁺ T cells (Fig. 5b–d and Supplementary Fig. 5). In addition, intracellular analysis showed that these effector cytokines were secreted by the same cells, illustrating polyfunctionality. However, the (polyfunctional) cytokine production of CD8⁺ T cells was dependent on cell composition, and only enhanced in AKT-inhibited cultures when starting with CD8⁺ T cells alone. Altogether, these data demonstrate that the beneficial effect of AKT-inhibition on the degranulation and polyfunctionality of CD8⁺ T cells is dependent on the composition of the cell starting material, where even though presence of CD4⁺ T cells in general improved (control) cultures, in these cultures they impede the favorable effect of AKT-inhibition on CD8⁺ T cells.

As most T cell therapies are generated using polyclonal stimulation, AKT-inhibition was also investigated in combination with T cell expansion upon anti-CD3/CD28 bead stimulation instead of allogeneic-DCs. For polyclonal expanded cells, the cell composition had less effect on memory differentiation and subset skewing than in allogeneic DC cultures (Fig. 6a–c). Moreover, the effects of AKT-inhibition on CD62L expression were less prominent, however could preserve expression of CCR7 and CXCR4 indicating an early memory phenotype of both CD4⁺ and CD8⁺ T cells (Fig. 6a, b). Similarly, also here a change in cytokine production was observed with less IFNγ-producing and more IL4-producing CD4⁺ T cells (Fig. 6c). In contrast to allogeneic-DC expanded cultures, as cell composition had minimal influence in these assays, the reduction in IFNγ-producing CD4⁺ T cells was also observed in CD3⁺ T cell cultures. Importantly, rechallenge experiments showed clearly enhanced expansion capacity of AKT-inhibited CD8⁺ T cells cultures, but only in single CD8⁺ T cell cultures (Fig. 6d). Similar to allogeneic DC cultures, this could not be observed in CD3⁺ T cell cultures were expansion of non-AKT-inhibited CD8⁺ T cells was already higher, most likely due to the help by CD4⁺ T cells (Fig. 6d). Combined, also in polyclonal-expanded T cell cultures, AKT-inhibition preserved the early memory phenotype of CD4⁺ and CD8⁺ T cells and skewed CD4⁺ T cells toward Th2 at the expense of a Th1 profile, with limited influence of cell composition. However, the rechallenge expansion capacity of CD8⁺ T cells in the presence of AKT-inhibited CD4⁺ T cells could not be further enhanced using AKT-inhibition.

Also for the polyclonal expanded CD8⁺ T cells, we analyzed degranulation and polyfunctional effector cytokine production after one week of rechallenge in absence of AKT-inhibitors. In contrast to allogeneic-expanded CD8⁺ T cells, no effect on degranulation capacity was seen for AKT-inhibition (Fig. 6e). Furthermore, while in allogeneic expanded cultures an increase was observed for cytokine production, here IL2 and TNFα production was not affected, and production of IFNγ was even reduced in polyclonal-expanded AKT-inhibited CD8⁺ T cells (Fig. 6f–l). Combined this resulted in a reduced polyfunctionality for AKT-inhibited CD8⁺ T cells. Altogether, these data demonstrate that the beneficial effect of AKT-inhibition on the degranulation and polyfunctionality of CD8⁺ T cells is dependent on the composition of the cell starting material and expansion strategy, where the presence of CD4⁺ T cells as well as polyclonal stimulation impede the favorable effect of AKT-inhibition on CD8⁺ T cells.