Restored immune cell functions upon clearance of senescence in the irradiated splenic environment

Abstract Some studies show eliminating senescent cells rejuvenate aged mice and attenuate deleterious effects of chemotherapy. Nevertheless, it remains unclear whether senescence affects immune cell function. We provide evidence that exposure of mice to ionizing radiation (IR) promotes the senescent‐associated secretory phenotype (SASP) and expression of p16INK4a in splenic cell populations. We observe splenic T cells exhibit a reduced proliferative response when cultured with allogenic cells in vitro and following viral infection in vivo. Using p16‐3MR mice that allow elimination of p16INK4a‐positive cells with exposure to ganciclovir, we show that impaired T‐cell proliferation is partially reversed, mechanistically dependent on p16INK4a expression and the SASP. Moreover, we found macrophages isolated from irradiated spleens to have a reduced phagocytosis activity in vitro, a defect also restored by the elimination of p16INK4a expression. Our results provide molecular insight on how senescence‐inducing IR promotes loss of immune cell fitness, which suggest senolytic drugs may improve immune cell function in aged and patients undergoing cancer treatment.

Substantial data demonstrate the impact of aging on the immune system promotes increased susceptibility to infection (Haynes & Swain, 2012;Kogut, Scholz, Cancro, & Cambier, 2012;Liu et al., 2011;Nikolich-Zugich, 2014). Several groups demonstrate aged individuals exhibit attenuated immune adaptive response, particularly reduced proliferation of T cells and dendritic cell function (You, Dong, Mann, Knight, & Yaqoob, 2014). Deletion of p16 INK4a in T cells enhanced antigen-specific immune response, which suggest senescence promotes an intrinsic defect in aged T cells (Liu et al., 2011). Moreover, mice lacking the expression of lamin A, a nuclear scaffolding protein, present an accelerated aging phenotype with immune deficiencies (Xin, Jiang, Kinder, Ertelt, & Way, 2015). In comparison, whether IR-induced senescence has a long-term impact on the immune system is less defined. A recent study showed mice irradiated (up to 4 Gy) at a young age failed to impair immune functions at old age (Pugh et al., 2016). Of note, the study compared irradiated aged mice (19 months) with F I G U R E 1 Exposure of mice to IR induces p16 INK4a expression and SASP in the spleen. (a) Schematic of the experimental design. Briefly, 12-week-old p16-3MR mice were exposed to 6.5 Gy total body irradiation, and 8 to 9 weeks later, mice were treated or not with GCV for 5 consecutive days to eliminate p16 INK4a -positive cells. (b) One day after the last GCV treatment, mice were injected i.p. with coelenterazine (CTZ), and 14 min later, mice were sacrificed. Spleens were surgically removed to quantify the luminescence. Representative photographs are shown. (c) Shown is the average integrated photon density emitted from p16-3MR mice exposed (+) or not (−) to IR and treated (+) or not (−) with GCV. (d) Quantification of endogenous p16 INK4a mRNA levels as determined by qPCR from full spleen lysates. 18S ribosomal RNAs was used as an internal control. (e) Expression levels of VEGF, IL-6, KC, MCP-1, IL-1α, and IL-10 from splenocyte lysates as detected by multiplex array. Shown is the median analyzed by one-way ANOVA ***p < 0.001; **p < 0.01; *p < 0.05; n = 5-8 mice per group age-matched nonirradiated counterparts which already exhibit diminished immune function.
We previously observed irradiated mice developed impaired lymphopoiesis in the bone marrow, an effect both cellular nonautonomous and dependent on p16 INK4a (Carbonneau et al., 2012). Our current study sought to investigate whether IR-induced p16 INK4a expression interfered with immune cell function. Using a previously described senescence mouse model (Demaria et al., 2014), we show that senescence-inducing IR impairs immune cell function in the splenic environment, an effect partially driven by the SASP and reversible with clearance of p16 INK4a -positive senescent cells.

| Exposure to IR induces features of senescence in the spleen
We previously showed that exposure to IR led to delayed (6-8 weeks) p16 INK4a expression in distinct mice tissues, including the spleen (Le et al., 2010;Palacio, Krishnan, Le, Sharpless, & Beausejour, 2017).
The cause for expression delay remains unclear but it protects mice against cancer progression (Palacio et al., 2017) To answer these questions, we first exposed mice to total body irradiation at a sublethal dose of 6.5 Gy, the maximum tolerated dose for mice to survive without requiring a bone marrow transplant ( Figure 1a). Due to the delay in p16 INK4 expression, we waited 8 weeks post irradiation to allow increase in p16 INK4a expression and restrict the time required to regain steady spleen cellularity. We could not detect one marker of senescence (senescence-associated β-galactosidase) in irradiated spleen tissue sections. However, we observed an increase in p16 inK4a and SASP factors by qPCR along with a decrease in lamin B1 expression in macrophages ( Figure S1A and B). Furthermore, we detected persistent DNA damage in stromal splenic cells, which did not appear in macrophages and hematopoietic cells ( Figure S1C and D). Of note, splenic cell counts never F I G U R E 2 Attrition of T-and B-cell populations in the irradiated spleen. (a, b) Shown are the purity (left panels) and quantification of p16 INK4a mRNA levels (right panels) of isolated B220 + and CD3 + cell populations as determined by flow cytometry and qPCR, respectively. 18S ribosomal RNA was used as an internal control. (c-e) Quantification by flow cytometry of the absolute cell counts for CD3 + CD4 + , CD3 + CD8 + , and B220 + populations per full spleen collected from mice treated as indicated. Cell counts were determined 1 day following the last injection of GCV. Shown is the average ± SEM. The ρ value was determined by a one-way ANOVA. *p < 0.05. n = 5-7 mice per group

| Attrition of CD3 + and B220 + cell populations in the irradiated spleen
A few days following IR, we detected substantial cell death in splenic T and B cells which also occurred in other cell populations to a lesser extent ( Figures S2 and S3). A result consistent with the fact that lymphocytes constitute a highly radiosensitive cell population. We asked whether IR induced p16 INK4a in lymphocytes and whether this impacted their absolute number after reconstitution. We collected spleens from control, irradiated mice untreated/treated with GCV and dissociated at the single-cell level. We isolated cell populations with magnetic columns (~90% purity as determined by flow cytometry) and extracted mRNA for qPCR analysis. We found p16 INK4a gene expression elevated fourfold in F I G U R E 3 IR impairs T-cell proliferation in vitro and in vivo. (a) Schematic of the experimental design. 12-week-old p16-3MR mice were exposed to 6.5 Gy total body irradiation, and 8-9 weeks later, mice were treated or not with GCV for 5 consecutive days. On day 6 (D6), mice were sacrificed and splenocytes were collected and labeled with CFSE. Alternatively, on day 15 (D15), mice were injected with the LCMV-Arm and sacrificed 1 week later. (b, c) The proportion of CD3 + cell undergoing proliferation following an allogenic stimulus was determined by flow cytometry. In panel b, the proliferation of T cells was determined from gated CFSE-CD3 + cells from p16-3MR splenocyte responder cells mixed with CD-1 stimulator splenocytes (ratio 1:2). In panel c, the proliferation of T cells was determined from gated CFSE-CD3 + from purified CD3 + responder cells mixed with CD-1 stimulator splenocytes (ratio 1:2). (d) Quantification of the capability of splenic T cells to proliferate in vivo following the injection of mice with the LCMV-Arm. Shown is the proportion of gp33 + CD8 + T cells undergoing LCMV-specific proliferation as determined by flow cytometry for the expression of the Ki67 proliferation marker. (e) Shown is the proportion of gp33 + CD8 + T cells expressing granzyme B (Grz) as determined by flow cytometry. For all graphs, the average ± SEM is shown from n = 5-7 mice except for panels d and e where each dot represents counts from an individual mouse. The ρ value was determined by a one-way ANOVA, ***p < 0.001; **p < 0.01; *p < 0.05 CD3 + cells and fivefold in B220 + cells (Figure 2a,b). These levels are similar to data from aged spleen-derived lymphocytes (Krishnamurthy et al., 2004). Absolute cell counts diminished approximately 50% 8 weeks post-IR for both CD3 + (CD4 + and CD8 + ) and B220 + cell populations ( Figure 2c-e). While GCV injections effectively eliminated p16 INK4a expressing cells in both populations, only the CD3 + CD4 + fraction exhibited cell levels observed in nonirradiated mice.

| Impaired proliferation of T cells depends on the irradiated splenic environment
We next explored the impact of p16 INK4a expression on the proliferative potential of irradiated splenic T cells. To address this question, we conducted a mixed lymphocyte reaction (MLR) with splenocytes from untreated or treated p16-3MR mice ( Figure 3a, left segment) The irradiated splenic environment impairs T cell proliferation. (a) Schematic of the experimental design. CD3 + effector cells were isolated by negative selection from the spleens of p16-3MR mice and labeled with CFSE. The splenic environment corresponds to CD3 + T cell-depleted splenocytes from p16-3MR mice previously (8-9 weeks) exposed (+) or not (−) to IR and treated or not with GCV. Stimulator cells were freshly irradiated (30 Gy) allogenic splenocytes collected from a CD-1 mouse. (b) Shown is the proportion of effector CD3 + cell undergoing proliferation in the presence of the indicated splenic environment following an allogenic stimulus (CD-1 stimulator). T cell proliferation was determined by flow cytometry (CFSE dilution). (c) Shown is the proportion of effector CD3 + cell undergoing proliferation following stimulation with anti-CD3/anti-CD28-coated beads in the lower well of a Transwell plate with the indicated (control or IR) splenic environment in the top well.  and freshly irradiated allogenic splenocytes obtained from a CD-1 mouse. A one-way MLR relies on the ability of "responder" cell (i.e., T cells within p16-3MR splenocytes) activation by allogenic HLA molecules displayed by "stimulator" cells (i.e., antigen-presenting cells in allogenic CD-1 splenocytes). The one-way MLR allowed us to restrict irradiated allogenic splenocytes to act as stimulators but not responders. Splenic T cells from p16-3MR mice irradiated 8-9 weeks earlier exhibit approximately 40% reduced proliferation compared to splenic T cells from nonirradiated mice (~65%-70% vs. ~35%-40%, see Figure 3b). Significantly, the proliferation of splenic T cells was restored in irradiated mice treated with GVC ( Figure 3b).
In contrast, magnetically purified T cells (not full splenocytes) did not demonstrate improved proliferative capacity (Figure 3c).
To confirm this result, we measured the capacity of splenic T cells to proliferate in vivo in response to an acute lymphocytic chori- Overall, these results demonstrate that IR does not impair intrinsic proliferation of splenic T cells. Irradiated splenocytes appear to compromise T-cell proliferation in vitro and in vivo which suggests the splenic environment, either cells or their secretome, interferes with T-cell proliferation. To test this notion, we performed a MLR with CD3 + effector T cells purified from p16-3MR mice and stimulator cells from freshly irradiated CD-1 splenocytes. We placed effector and stimulator cells in the presence of an "environment" of CD3 + -depleted splenocytes isolated from p16-3MR mice (Figure 4a).
This experimental setup confirmed that the presence of a previously (8-9 weeks) irradiated splenic environment compromised the proliferative capacity of purified T cells (Figure 4b). Additionally, we partially restored T cell proliferation when we used splenic environment from GCV-treated mice with GCV (Figure 4b).
To identify whether irradiated splenocytes or their secretome proved detrimental to T cell proliferation, we performed a modified MLR whereby we separated effector/stimulator cells from splenocytes using a transwell. Thus, preventing cellular interaction allowed us to determine whether the splenic environment (secretome) effectively interfered with proliferation. Using beads coupled to anti-CD3 and anti-CD28 antibodies as stimulator, we found that delay in T cell proliferation only required exposure to irradiated splenocyte secretome (Figure 4c,d). We confirmed the negative impact of irradiated secretome on T cell proliferation, testing allogenic splenocytes F I G U R E 5 Impaired macrophage and DC counts and function in the splenic environment. (a, b) Shown are the purity (left panels) and quantification of p16 INK4a mRNA levels (right panels) of isolated F4/80 + macrophages and CD11c + DC cell populations as determined by flow cytometry and qPCR, respectively. 18S ribosomal RNA used as an internal control. (c, d) Shown is the quantification by flow cytometry of the absolute cell counts per spleens for F4/80 + and CD11c + cell populations, respectively, collected from mice treated as indicated. Cell counts were determined 1 day following the last injection of GCV. (e, f) Quantification of the proportion of purified F4/80 + macrophages and CD11c + DC populations capable of phagocytosis. Shown is the average ± SEM from n = 7-15 mice per group except for panels e and f where each dot represents counts from an individual mouse. The ρ value was determined by a one-way ANOVA. ***p < 0.001; **p < 0.01; *p < 0.05

| IR impairs macrophages and dendritic cells in the spleen
Macrophages and dendritic cells (DC) play a pivotal role in innate and adaptive immunity through their capacity to phagocyte, process, and present antigens. These cells exhibit more radioresistance than lymphocytes 1 week following exposure to IR ( Figure   S2 and S3). However, 8-9 weeks after IR, these cells (  Whether a decrease in phagocytosis is sufficient to negatively impact T cell proliferation in the context of a MLR requires more study.

| D ISCUSS I ON
We also observe a slight increase in the number of macrophages following the injection of GCV in irradiated mice, an unexpected result given the high level of p16 INK4a expression in this population.
We speculate the increase may result from peripheral macrophages recruited to the spleen in response to GCV-induced cell death.
Recent data demonstrate immune stimuli modulate the expression of p16 INK4a in macrophages, suggesting a complex regulation in these cells and more work required to delineate complex regulation (Hall et al., 2017).
Studies in aging mice and humans show an increase in the proportion of regulatory T cells (T regs) that impair immune function (Sharma, Dominguez, & Lustgarten, 2006;Simone, Zicca, & Saverino, 2008). We quantified the proportion of T regs in the spleen and found their number sharply decreased in irradiated mice (with and without treatment with GCV), suggesting these cells are unlikely to negatively impact proliferation of T cells ( Figure S5). Finally, IR may act on the stromal architecture of the spleen stroma. Indeed, we observe a significant increase in p16 INK4a expression and decrease in absolute cell counts following the injection of GCV in sub-populations (gp38 + and CD35 + ) of splenic stromal cells ( Figure S6). Hence, while we show that the SASP negatively affects proliferation of splenic T cells, the overall impact of IR on the spleen function may alter multiple signaling pathways and is likely multifactorial.
In conclusion, we demonstrate that elimination of p16 INK4a

| Animals and treatments
p16-3MR mice were kindly donated by Dr. Judith Campisi (Buck Institute) and breed on site according to a material transfer agreement (Demaria et al., 2014). All in vivo manipulations were approved by the Comité Institutionnel des Bonnes Pratiques Animales en Recherche of the CHU Ste-Justine. 12-to 14-week-old p16 INK4a -3MR mice were exposed to X-rays at the single sublethal dose of 6.5 Gy (1 Gy/min) using a Faxitron CP-160. GCV was administrated daily by intraperitoneal (i.p.) injections for 5 consecutive days at a dose of 25 mg/kg in 1X-PBS (Sigma).

| Bioluminescence
To detect luminescence from the 3MR gene cassette, mice were

| Flow cytometric analysis
To obtain absolute cell counts from various populations, spleens were

| Multiplex cytokine analysis
The splenic cell secretome was quantified by Eve Technologies protease inhibitors (PI). Lysates were centrifuged at 10,000 g for 10 min at 4°C and supernatants transferred to a new tube and normalized with 1X-PBS to 0.5 mg/ml of proteins.

| T cell proliferation assays
T cell proliferation was evaluated by an allogenic mixed lymphocyte reaction (MLR) assay. Briefly, cells were harvested from spleens of p16-3MR mice irradiated 8-9 weeks earlier and untreated or treated with GCV for 5 consecutive days prior to sacrifice. Splenocytes or purified CD3 + cells were labeled with CellTrace™ 6-carboxy-succinimidyl-fluorescein-ester dye (CFSE) and used as responder. CD3 + cells were isolated by negative selection using the EasySep Mouse T Cell Isolation Kit (Catalog, 19851, STEMCELL). The purity of CD3 + cells was determined by flow cytometry using a PE-conjugated anti-CD3 antibody (Catalog, 100240, BioLegend). Splenocytes obtained from the outbred CD-1 ® IGS mouse strain (Charles River) were used as stimulator and irradiated at a dose of 30 Gy. Mixed lymphocyte reactions were set up with 1 × 10 5 CFSE-labeled p16-3MR responder cells and 2 × 10 5 freshly irradiated allogenic CD-1 stimulator splenocytes in round-bottom 96-well plates, at 37°C, 5% CO 2 for 3 days.

| Statistical analysis
GraphPad Prism 7 software was used for statistical analysis; ρ values on multiple comparisons were calculated using one-way analysis of variance (ANOVA) with Bonferroni post hoc test.

ACK N OWLED G M ENTS
We and H.D. are supported by scientist awards from the Fonds de la recherche du Québec-Santé.

CO N FLI C T O F I NTE R E S T
None declared.