p16^Ink4a, a marker of cellular senescence, is associated with renal disease in the B6.NZMSle1/Sle2/Sle3 mouse model of lupus

Mice

B6 and B6.Sle1.2.3 mice were bred and maintained in the SPF animal facility of the university. Plasma (from retro-orbital blood), urine and kidneys (cut in half longitudinally, for formalin-fixed paraffin-embedded (FFPE) and optimal cutting temperature (OCT)-embedded blocks) were collected from two cohorts of mice. The first consisted of 17 female B6.Sle1.2.3: 4 mice of 3 months of age and 13 mice of 9–11 months of age. The aged mice were selected based on proteinuria test strips (Combur Test E, Roche): 6 mice with no or low proteinuria (<2) and 7 mice with high proteinuria (≥2). The second (time course) cohort consisted of 59 mice: 2 B6 mice (1 male and 1 female) and 8 B6.Sle1.2.3 mice (4 male and 4 female, except at 10 months: 4 male and 3 female) each, sacrificed every 2 months from 2 to 12 months of age.

ELISA and colorimetric assays

Measurements of total IgG (Invitrogen, 88-50400) and anti-dsDNA IgG (Shibayagi, ARKDD-061) were performed according to manufacturers’ protocols, on plasma diluted to fit onto standard curves. Results are expressed in μg/mL and U/mL, respectively. Albumin concentration in urine was measured using the Mouse Albumin Matched Antibody Pair kit (Abcam ab210890). Briefly, 96-well high-binding plates were coated overnight at 4°C with 2 µg/mL capture antibodies in a 35 mM NaHCO₃, 15 mM Na₂CO₃, pH 9.6 solution and blocked (1% BSA, 0.05% Tween 20 in PBS) for 2 hours at room temperature. Next, mouse albumin standards (from 12.5 ng/mL to 0.195 ng/mL) and urine samples (diluted in blocking buffer) were incubated on a shaker at 400 rpm for 2 hours at room temperature. Detection antibodies (0.5 µg/mL) were then added for 1 hour with shaking at room temperature, followed by 0.05 µg/mL Streptavidin-HRP. Finally, TMB solution was added and the reaction stopped with H₂SO₄ 1N. Plates were washed three times with PBS-0.05% Tween 20 between each step. Optical density was measured at 490 nm with subtraction of background at 560 nm. Results are expressed in mg/dL. A mouse creatinine ELISA kit (Abcam, ab65340) was used on urine diluted to fit on the standard curve, following the manufacturer’s protocol. Results are expressed in g/dL. Proteinuria was calculated as the urine albumin/creatinine ratio, expressed in mg of albumin per g of creatinine.

Immunofluorescence

IgG deposition was detected on 5 µm OCT serial kidney sections, with 1 µg/mL anti-mouse IgG antibody coupled to AlexaFluor 488 (ThermoFisher, A-11001). Counterstaining was performed using 1 µg/mL Hoechst 33342. Slides were digitalised on a Pannoramic Confocal scanner (3DHistech) at ×20 magnification. Glomeruli were scored from 0 (no IgG deposit) to 3 (severe IgG deposit). Values shown are the mean scores for 30 glomeruli from each kidney section, from three blinded scorers.

Senescence-associated β-galactosidase assay

Senescence-associated β-galactosidase (SA β-gal) assay was performed on 15 µm OCT serial kidney sections. Briefly, slides were fixed in 0.2% formaldehyde for 10 min. After washing with PBS, slides were incubated with the staining solution (40 mM citric acid/Na phosphate pH 5.7, 5 mM K₄[Fe(CN)₆].3H₂O, 5 mM K₃[Fe(CN)₆], 150 mM NaCl, 2 mM MgCl₂, 1 mg/mL X-gal (Sigma-Aldrich, B4252)) overnight at 37°C in an incubator without CO₂. Slides were then washed and counterstained 30 s with haematoxylin. Slides were digitalised on a Pannoramic Scan II scanner (3DHistech) at ×20 magnification. A semi-quantitative score was given for 10–20 glomeruli per kidney section, with each glomerulus classified as moderate (1), intermediate (2) or high (3).

Histology and immunohistochemistry

Immunostaining with 1 µg/mL anti-CD8 (Cell signalling, 98941S) and anti-p16^Ink4a (Abcam, ab211542), Picrosirius red (PSR) and H&E staining were performed on 5 µm FFPE serial kidney sections. PSR staining was performed as previously described.34 Slides were digitalised on a Pannoramic Scan II scanner (3DHistech) at ×20 magnification. Computer-assisted quantification of p16^Ink4a-positive and CD8-positive cells over entire sections was performed using Author V.2017.2 (Visiopharm). Results shown are the number of p16^Ink4a-positive or CD8-positive cells per mm² of tissue. Semi-quantitative PSR staining scores (scale: 0–6) were obtained by adding the median glomerular fibrosis score and the median interstitial fibrosis score (absent: 0, low: 1, intermediate: 2, high: 3), from three blinded scorers. Activity and chronicity scores are median scores from three blinded scorers, based on H&E-stained kidney sections. Activity scores evaluate endocapillary proliferation, extracapillary proliferation, periglomerular and interstitial immune cell infiltration (each parameter scored between 0 and 3, total scale: 0–12). Chronicity scores evaluate glomerulosclerosis, thickness of fibrous tissue lining Bowman’s capsules, tubular atrophy and interstitial fibrosis (each parameter scored between 0 and 3, total scale: 0–12).

Statistical analyses

Statistical analyses were performed on GraphPad Prism V.9.1.0: Mann-Whitney rank test for two-group comparisons, Kruskal-Wallis with Dunn’s post-hoc tests for multiple-group comparisons, nested t-test for comparisons of subgroups within two groups and Spearman’s rank-order correlation coefficient. Penetrance of disease parameters was calculated as: the percentage of mice displaying values above the mean plus 2 SD of values from all 12 (2–12 month old) B6 mice.

Cells positive for p16^Ink4a, a classic marker of cellular senescence, are observed in kidneys of aged B6.Sle1.2.3 lupus-prone mice

We conducted an exploratory study on a cohort of (n=17) female B6.Sle1.2.3 mice, to assess for the presence of p16^Ink4a-positive cells in the kidney in this lupus-prone strain. Four mice of 3 months of age were included as pre-symptomatic controls; 13 mice of 9–11 months of age were selected based on urine protein estimated by dipstick, to capture kidney disease heterogeneity: 6 mice with low proteinuria (values: 0–1) and 7 with high proteinuria (values: ≥2). Quantitative measurement of proteinuria confirmed dipstick-based selection of the cohort: urine albumin/creatinine ratio was low in young mice (14.8–26.7 mg/g) and highly variable among aged mice (31.4–42 702.1 mg/g) (online supplemental figure 1A), with a significant difference between the low (0–1) and high (2–3) dipstick groups (online supplemental figure 1B). Histopathological (activity and chronicity) scores also differed significantly between the low and high dipstick groups (online supplemental figure 1C,D), as did glomerular IgG deposition (online supplemental figure 1E). In contrast to kidney disease, systemic disease parameters (total IgG and anti-dsDNA antibody levels in plasma) were similar between the groups (online supplemental figure 1F,G), suggesting greater homogeneity of systemic autoimmunity.

As observed in samples from patient with LN,7 p16^Ink4a protein expression was highly variable among older mice: few p16^Ink4a-positive cells were detected in kidney sections of certain mice (illustrated in figure 1A, top left), comparable with kidneys from young mice (data not shown); others showed strong p16^Ink4a staining (illustrated in figure 1A, top right). Quantification of p16^Ink4a staining (p16^Ink4a-positive cells/mm²) confirmed this heterogeneity (range: 5.9–186.2, median: 42.2) (figure 1B). p16^Ink4a-positive cells were mainly observed in glomeruli, with staining of the glomerular tuft (containing endothelial cells, podocytes and mesangial cells) and strong positivity in Bowman’s capsules (parietal epithelial cells) (figure 1C). Tubular cells and scattered cells in the tubulointerstitial compartment were also positively stained (figure 1C): a pattern similar to that observed in kidney biopsies from patients with LN.7 We performed an SA β-gal assay (a second widely used marker of cellular senescence) on frozen sections from low and high p16^Ink4a-positive kidneys (≤50th vs >50th percentile, n=3 each). Semi-quantitative evaluation of glomerular β-gal staining showed that kidneys characterised by high p16^Ink4a displayed significantly higher glomerular β-gal scores than low p16^Ink4a kidneys (figure 1D,E). This suggests that cellular senescence is indeed detected in a proportion of aged B6.Sle1.2.3 mouse kidneys.

Request permission

The presence of p16^Ink4a-positive cells is associated with renal disease severity in B6.Sle1.2.3 mice

We next asked whether p16^Ink4a is associated with kidney disease, assessed by several parameters: urine albumin/creatinine ratio, histopathology (activity, chronicity) scores and glomerular IgG deposits in kidney sections. Mice with p16^Ink4a-high kidneys (upper 50th percentile: >Q2) displayed significantly higher urine albumin/creatinine ratios (figure 2A), as well as higher activity and chronicity scores (figure 2B,C). Accordingly, a significant positive correlation was observed between p16^Ink4a staining and urine albumin/creatinine ratios and activity and chronicity scores (figure 3A–C). Conversely, significantly lower glomerular IgG deposit scores were observed in p16^Ink4a-high kidneys (figure 2D); accordingly, a significant negative correlation was observed between these values (figure 3D). This may reflect greater activation of renal cells, immune cell recruitment and (complement and Fc receptor-mediated) immune complex clearance leading, in fine, to greater tissue damage in mice with advanced kidney disease as compared with mice with mild kidney disease. In contrast to its correlation with kidney disease severity, p16^Ink4a was not associated with parameters of systemic disease such as total IgG and anti-dsDNA antibody levels in plasma (figure 2E,F).

Request permission

Request permission

It has been shown that CD8⁺ T lymphocytes are the predominant immune cell type infiltrating the LN kidney and that their presence is associated with renal disease severity.35 36 We previously showed a significant association between CD8⁺ T cell infiltration, fibrosis and p16^INK4a-positive cells in kidney biopsies of patients with LN.7 Sections (in series with those stained for p16^Ink4a) of mouse kidney were therefore stained for CD8⁺ cells and with PSR for collagen fibres (figure 1A). Quantification showed that p16^Ink4a-high biopsies displayed significantly higher CD8⁺ T cell infiltration (figure 2G); they also showed significantly higher collagen deposition (with glomerulosclerosis, thickening of Bowman’s capsules and interstitial fibrosis) as reflected by semi-quantitative scores of PSR staining (figure 2H). Accordingly, p16^Ink4a staining correlated positively with CD8⁺ T cell infiltration and with fibrosis (figure 3E,F).

Accumulation of p16^Ink4a-positive cells in B6.Sle1.2.3 kidneys occurs at 8 months of age, later than systemic disease and overlapping with renal disease

In order to study the temporal relationship between the appearance of p16^Ink4a, and systemic and renal disease parameters, we performed a time course in ageing mice. Two B6 controls and 8 B6.Sle1.2.3 lupus-prone mice were sacrificed at 2-month intervals, from 2 to 12 months of age. Penetrance of disease parameters (the percentage of B6.Sle1.2.3 mice displaying values above the mean plus 2 SD of all 12 B6 mice) at the different time points is displayed in online supplemental figure 2. Systemic disease indices as well as glomerular IgG deposits appeared in B6.Sle1.2.3 mice at 2–4 months of age (figure 4A–C and online supplemental figure 2A,B). Kidney disease became apparent later, with proteinuria (urine albumin/creatinine ratio) and histopathological scores rising from 6 months of age (figure 4D–F and online supplemental figure 2B). The appearance of p16^Ink4a-positive cells, CD8⁺ T cell infiltration and collagen deposition seemed to occur concurrently, at 8 months of age (figure 4G–I and online supplemental figure 2C). High penetrance (>50%) of glomerular IgG deposits and systemic disease parameters was observed at 4 and 6 months of age, respectively. In contrast, high penetrance of renal disease parameters (activity and chronicity scores, urine albumin/creatinine ratio) and p16^Ink4a occurred later, at 8 months of age. Finally, fibrosis (PSR staining) affected 50% of B6.Sle1.2.3 mice only by 12 months of age, while the peak penetrance of CD8⁺ T cell infiltration (at 8 months of age) was only 25% in B6.Sle1.2.3 mice (online supplemental figure 2C).

Request permission

In order to assess for whether p16^Ink4a, once it appeared at 8 months of age, was indeed associated with kidney disease in this second cohort, we pooled data from the mice 8–12 months old. Aged B6.Sle1.2.3 mice in the upper 50th percentile (>Q2) for p16^Ink4a staining showed significantly higher renal disease scores, CD8⁺ T cells and PSR scores, than those in the lower 50th percentile (≤Q2; figure 5A–E). Indeed, the ≤Q2 p16^Ink4a B6.Sle1.2.3 group did not differ significantly from B6 on renal disease parameters and fibrosis, whereas the >Q2 p16^Ink4a group did (figure 5A–E). Glomerular IgG deposits did not differ significantly between groups (figure 5F). Finally, systemic disease scores (total IgG and dsDNA antibody levels in plasma) did not differ between the ≤Q2 and >Q2 p16^Ink4a B6.Sle1.2.3 groups, and were higher in both as compared with B6 (figure 5G,H). Together, this suggests that in the context of (relatively consistent) systemic autoimmune insult to the kidney, the cellular senescence marker p16^Ink4a may be useful in flagging (highly heterogeneous) renal disease in B6.Sle1.2.3 mice.

Request permission