C57BL/6 mice were used, maintained in the Animal Housing Facility of the KU Leuven under conditions of constant temperature, humidity and 12-hour light-dark cycle, with ad libitum access to water and food. Animal experiments were approved by the KU Leuven Ethical Committee for Animal Experimentation (P153/2018).

Mice heterozygous for the IL-6 mutation (Il6^-/+ (Il6^tm1Kopf )), a neo^r cassette insertion in the first coding exon (17), were crossed to obtain homozygous offspring (Il6^+/+ or Il6^-/- ) which were analyzed at embryonic stages E12.5 and E16.5 (E0.5 as day of vaginal plug), and neonatal [postnatal day (PD) 7], (young-)adult [8-12 weeks (wks)] or middle-aged [10-15 months (mo)] phases of life. Offspring was genotyped for the presence of the neo^r cassette by PCR using 5’-TTCCATCCAGTTGCCTTCTTGG-3’ as common forward primer, 5’-TTCTCATTTCCACGATTTCCCAG-3’ as wildtype (WT) reverse primer and 5’-CCGGAGAACCTGCGTGCAATCC-3’ as mutant reverse primer.

Gh^Cre/+ (Tg(Gh1-cre)bKnmn) mice were crossed with ROSA26^iDTR/iDTR mice (Gt(ROSA)26Sor^{tm1(HBEGF)Awai} ) to create Gh^Cre/+;ROSA26^iDTR/+ offspring (5), which was genotyped for the presence of the Cre transgene by PCR using 5’-TGCCACGACCAAGTGACAGCAATG-3’ as forward and 5’-ACCAGAGACGGAAATCCATCGCTC-3’ as reverse primer. These mice were further crossed to the Il6^-/- genotype to create triple transgenic offspring, i.e.

Gh^Cre/+;ROSA26^iDTR/+;Il6^-/- and associated genotypes Gh^Cre/+;ROSA26^iDTR/+;Il6^+/+ , Gh^+/+;ROSA26^iDTR/+;Il6^+/+ and Gh^+/+;ROSA26^iDTR/+/Il6^-/- . The mice were intraperitoneally (i.p.) injected with diphtheria toxin (DT, 4 ng/g body weight; Sigma-Aldrich) twice a day for 3 consecutive days. Pituitaries, more in particular the major endocrine anterior lobes (ALs), were isolated and analyzed the day after DT injection (day (d) 4), one week later (d11) or 5 mo later (5).

Pregnant mice were euthanized using CO₂ asphyxiation followed by cervical dislocation. Individual embryos were isolated from the uterus and extra-embryonic membranes removed.

For isolation of the postnatal pituitary, mice were euthanized using CO₂ asphyxiation followed by decapitation. The intact pituitary was isolated for immunostaining analysis, while the AL was separated from the intermediate lobe (InL) and posterior lobe (PL) under the stereomicroscope for dissociation into single cells with trypsin as previously described (12), to be followed by gene expression and quantitative immunofluorescence analyses.

Embryos and intact pituitaries were fixed with 4% paraformaldehyde (PFA, 4% in PBS; Merck) for 24 h at 4°C and 1-3 h at room temperature (RT), respectively. Samples were dehydrated with the Excelsior ES Tissue Processor (Thermo Fisher Scientific) and embedded in paraffin using the Histostar Embedding Workstation (Thermo Fisher Scientific). Paraffin sections (5 µm), in sagittal plane for embryos and coronal plane for pituitary samples, were made with the Microm HM360 (Thermo Fisher Scientific). Sections were dewaxed, rehydrated and subjected to antigen retrieval using citrate buffer (pH 6; 30 min at 95°C). Samples were permeabilized with 0.1% Triton X-100 (in PBS; Sigma-Aldrich), incubated with blocking buffer (0.15% glycine (VWR chemicals BDH), 2 mg/mL BSA (Serva), 0.1% Triton X-100 in PBS) with 10% donkey serum (Sigma-Aldrich) for 1 h at RT, and then with primary antibodies ( Supplementary Table 1 ) overnight at 4°C. Secondary antibodies ( Supplementary Table 1 ) and nuclear dye Hoechst33342 (2 μg/mL; Sigma-Aldrich) were added for 1 h at RT. Sections were mounted with ProLong Gold (Thermo Fisher Scientific) and images acquired using a Leica DM5500 upright epifluorescence microscope (Leica).

For quantitative immunofluorescence analysis, dissociated AL cells were spun down onto SuperFrost glass slides (Thermo Fisher Scientific) using the Shandon CytoSpin 3 Cytocentrifuge (20,000-50,000 cells/slide; 800 rpm; 10 min). Cytospin samples were subsequently dried for 10 min, followed by a 15 min fixation in 4% PFA at RT. Cells were permeabilized with saponin (0.25% in PBS; Sigma-Aldrich) for 15 min and then blocked with donkey serum (10% in 0.25% saponin/PBS) for 20 min, both steps performed at RT. Primary antibodies were added overnight at 4°C ( Supplementary Table 1 ). Cells were labelled with secondary antibodies ( Supplementary Table 1 ) and nuclear dye Hoechst33342 for 1.5 h at RT, and sections mounted with ProLong Gold. Pictures were taken with a Leica DM5500 upright epifluorescence microscope. Ratios of immunoreactive cells were quantified in at least 10 random fields per slide. A range of 300-1500 cells was counted per slide using the cell counter plugin in Fiji software (18), with 2-4 slides analyzed per condition for each biological replicate. Considering the significant reduction in total cell number in the damaged AL, as found at final ablation (d11), we calculated the absolute number of the different immune-positive cells for comparisons at d11, using their proportion and the total number of cells obtained per AL, as described before (5).

Total RNA of embryonic pituitary, and of neonatal and adult dissociated AL cells was isolated using the RNeasy Micro kit (Qiagen) and subjected to reverse transcription (RT) with Superscript III First-Strand Synthesis Supermix (Invitrogen) according to the manufacturers’ protocol. SYBR Green-based RT-quantitative PCR (RT-qPCR) was applied on the cDNA samples using the StepOnePlus Real-Time PCR System (AB Applied Biosystems) and the Platinum SYBR Green qPCR Supermix-UDG (Thermo Fisher Scientific). Forward and reverse primers ( Supplementary Table 2 ) were designed with PrimerBank (19) and PrimerBLAST (20). β-actin (Actb) was used as housekeeping gene for normalization, and each sample was run in duplicate. Relative gene expression levels were calculated as ΔCt values (Ct_target – Ct_{housekeeping gene}). Gene expression levels were compared between sample and reference as relative expression ratio (fold change), calculated by the formula 2^{-(ΔCt,sample – ΔCt,reference)}, or as log₂ transformation of the fold change.

Statistical analysis was performed (when n ≥ 3) using Graphpad Prism (v9.4.1; GraphPad Software). Statistical significance was defined as P < 0.05.

We assessed the embryonically developing pituitary in IL-6 KO (Il6^-/- ) as compared to WT (Il6^+/+ ) mice at two time points, i.e. E12.5 when the pituitary primordium Rathke’s pouch (RP) is definitively formed, and E16.5 when the different endocrine lineages [corticotrope (ACTH+), gonadotrope [glycoprotein a-subunit (aGSU)+], and somatotrope/lactotrope/thyrotrope (PIT1+)] have emerged ( Figure 1A ) (21). Of note, Il6 is expressed in the embryonic pituitary, although at lower levels than in the adult gland ( Supplementary Figure 1 ). The morphology of the developing pituitary appeared not different between Il6^+/+ and Il6^-/- embryos ( Figures 1B-D ). In addition, the embryonic stem/progenitor cells, located around RP lumen (cleft) and marked by SOX2, as well as by the other pituitary stem cell markers E-cadherin (E-CAD) and cytokeratin (CK) 8 and 18 (11, 22), did also not display visible differences ( Figure 1B ). Similarly, the proliferative cell landscape, as visualized using the proliferation marker Ki67 and being most pronounced in the stem/progenitor cell zone around RP cleft, remained comparable between the Il6^-/- and Il6^+/+ developing gland ( Figure 1C ). Finally, the lack of IL-6 did not discernibly affect the emergence and development of the endocrine lineages (αGSU⁺, ACTH⁺, PIT1⁺) ( Figure 1D ).

Together, IL-6 does not play an indispensable or decisive role in pituitary organogenesis, a stage during which embryonic stem/progenitor cells actively give rise to the emerging endocrine cells.

In the first weeks after birth, the mouse pituitary undergoes a dynamic growth and maturation process while containing an activated stem cell population (10, 13, 23–25). Here, we investigated whether IL-6 is involved in this active neonatal pituitary remodeling. Il6 is expressed in the neonatal (PD7) gland, but still lower than in the later adult stage ( Supplementary Figure 1 ) (10). The stem cell compartment (encompassing the marginal zone lining the residual RP cleft), as marked by SOX2, E-CAD and CK8/18, did not show overt differences between Il6^+/+ and Il6^-/- neonatal pituitary ( Figure 2A ), in line with our recent findings that the number of SOX2⁺ cells and their proliferating subfraction does not change in IL-6 KO neonatal pituitary (10). Similarly, the topography and size of the different endocrine cell populations (i.e. ACTH⁺, PRL⁺, αGSU⁺, GH⁺) was not affected in the Il6-deficient neonatal pituitary ( Figures 2B, C ), robustly substantiating and expanding our recent findings of unaffected gene expression of lineage progenitor and endocrine cell markers (10). In addition, proliferation within the individual hormonal cell lineages did also not significantly change in Il6^-/- versus Il6^+/+ condition ( Figures 2B, C ).

Taken together, IL-6 is found to be dispensable for the pituitary’s neonatal maturation process encompassing activated stem cells and expanding endocrine lineages.

Then, we assessed the impact of IL-6 absence on the pituitary stem and endocrine cell phenotype of (young-)adult (8-12 weeks-old) and aging (10-15 months-old) mice, in which Il6 is prominently expressed [(7) and Supplementary Figure 1 ]. Lack of IL-6 did not visibly impact the expression pattern of the stem cell markers SOX2, E-CAD and CK8/18 in both adult and aging gland ( Figure 3A ), neither is there an overt (significant) change in the proportion of SOX2⁺ cells or their proliferative subfraction ( Figure 3B ), findings that further expand and corroborate our recent observations as analyzed for adult mice (7). Furthermore, absence of IL-6 did generally not entail significant changes in the endocrine cell lineages or proliferating (Ki67⁺) cells, except for a small increase in corticotropes (ACTH⁺ cells) and in proliferative cells in the aging Il6^-/- pituitary ( Figure 3C ), of which the (biological) relevance is presently unclear (see Discussion).

Taken together, IL-6 is not essential in the homeostatic process of adult and aging pituitary, during which cell turnover is very limited and stem cells are highly quiescent (2, 3, 7, 13).

Following local injury in the adult pituitary, as inflicted by somatotrope cell ablation using the Gh^Cre/+;ROSA26^iDTR/+ mouse model, the quiescent stem cells become promptly activated and enhance in proliferative activity (5–7, 26). Here, damage was inflicted in triple transgenic Gh^Cre/+;ROSA26^iDTR/+;Il6^-/- mice (further referred to as damaged (DMG)–KO) by DT injection for 3 consecutive days, and the pituitary compared to the appropriate DT-injected controls [Gh^+/+;ROSA26^iDTR/+;Il6^+/+ (CTRL-WT), Gh^Cre/+;ROSA26^iDTR/+;Il6^+/+ (DMG-WT), Gh^+/+;ROSA26^iDTR/+;Il6^-/- (CTRL-KO)] ( Figure 4A ). We observed a complete absence of the acute (d4) proliferative stem cell reaction (i.e. increase in SOX2⁺/Ki67⁺ cells as found in DMG-WT) when IL-6 is lacking ( Figure 4B ). This finding is in line with, but much more pronounced than, the 50% reduction in proliferative SOX2⁺ cell response that we recently reported through pharmacological IL-6 inhibition by in vivo antibody administration (7).

In previous work, we discovered that the adult pituitary possesses regenerative competence upon injury, as demonstrated by substantial (~50%) regeneration of the somatotrope cell population, observed 4-6 months after their obliteration (5, 6). Intriguingly, despite the lack of an acute stem cell reaction in the absence of IL-6, we here still detected significant somatotrope regeneration levels (~40%) in the DMG-KO pituitary when analyzed 5 months after the somatotrope ablation ( Figures 4A, C ), reaching a GH⁺ cell proportion similar to that in DMG-WT gland ( Supplementary Figure 2A ), thereby indicating comparable regeneration grade. The finding of regenerative realization is not due to lower extent of damage (ablation grade) inflicted in DMG-KO pituitary, which is indeed comparable (~80%; Figure 4C ) to the grade as reported before in the presence of IL-6 (5, 6). As another possible explanation, we examined whether other cytokines of the IL-6 family, or related factors, may rescue the function of IL-6 after pituitary damage in the IL-6 KO mouse. Expression levels of these cytokines in basal undamaged pituitary were not different between Il6^+/+ and Il6^-/- AL, except for Il22 which was lower in the IL-6 KO AL ( Supplementary Figure 2B ). Interestingly, the expression of Il22, as well as of Il11, substantially increased immediately after pituitary damage (d4) in the IL-6 WT animals, which quickly returned to baseline levels one week later (d11; DMG-WT in Figure 4D ). Of note, expression of IL-6, also upregulated immediately after damage, remained higher at d11 ( Supplementary Figure 2C ). Il22 and Il11 expression was also increased in the absence of IL-6 (DMG-KO; Figure 4D ) although at a significantly lower level than in DMG-WT ( Figure 4D ). Intriguingly, levels did not return to control one week later but remained elevated ( Figure 4D ). A similar pattern of sustained expression, different from control, was found for Lif but not for the other cytokines tested (Cntf, Osm, Ifng) ( Supplementary Figure 2D ). Finally, the number of SOX2⁺Ki67⁺ cells in the pituitary stem cell compartment showed a slight trend of increase, although not significant, in DMG-KO versus CTRL-KO pituitary at this later timepoint (d11; Figure 4E ). Together, these findings advance the hypothesis that, in the absence of IL-6, the lower but prolonged upregulation of other cytokines may compensate for the lack of IL-6 and might still generate a stem cell reaction although in a delayed and less forceful manner. However, it should be noted that this interpretation is still hypothetical and should be taken with the necessary caution (see Discussion).

In conclusion, IL-6 plays a key role in the acute pituitary stem cell activation response upon local damage. In its absence, other cytokines such as IL-11 and IL-22 may compensate, with lower but prolonged expression, thereby potentially entailing a delayed (and less potent) stem cell activation response toward eventual regenerative realization (summarized as hypothetical model in Figure 5 ).