Paradoxical effects of all-trans-retinoic acid on lupus-like disease in the MRL/lpr mouse model

Abstract

Roles of all-trans-retinoic acid (tRA), a metabolite of vitamin A (VA), in both tolerogenic and immunogenic responses are documented. However, how tRA affects the development of systemic autoimmunity is poorly understood. Here we demonstrate that tRA have paradoxical effects on the development of autoimmune lupus in the MRL/lpr mouse model. We administered, orally, tRA or VA mixed with 10% of tRA (referred to as VARA) to female mice starting from 6 weeks of age. At this age, the mice do not exhibit overt clinical signs of lupus. However, the immunogenic environment preceding disease onset has been established as evidenced by an increase of total IgM/IgG in the plasma and expansion of lymphocytes and dendritic cells in secondary lymphoid organs. After 8 weeks of tRA, but not VARA treatment, significantly higher pathological scores in the skin, brain and lung were observed. These were accompanied by a marked increase in B-cell responses that included autoantibody production and enhanced expression of plasma cell-promoting cytokines. Paradoxically, the number of lymphocytes in the mesenteric lymph node decreased with tRA that led to significantly reduced lymphadenopathy. In addition, tRA differentially affected renal pathology, increasing leukocyte infiltration of renal tubulointerstitium while restoring the size of glomeruli in the kidney cortex. In contrast, minimal induction of inflammation with tRA in the absence of an immunogenic environment in the control mice was observed. Altogether, our results suggest that under a predisposed immunogenic environment in autoimmune lupus, tRA may decrease inflammation in some organs while generating more severe disease in others.

Fig 1. Immunogenic environment in 6-week-old MRL/lpr mice.

(A) Anti-dsDNA IgG in the plasma of 6-week-, 9-week- and 17-week-old MRL and MRL/lpr mice as detected by ELISA. Data shown as relative levels (RL) were normalized to the average absorbance value of MRL mice at the same age, which was defined as 1. (B) Total IgG and IgM concentrations in the plasma of 6-week-old mice as detected by ELISA. (C-I) The percentages and absolute numbers of T, B, and dendritic cells in the spleen, mesenteric lymph node (MLN), and bone marrow (BM) of 6-week-old mice as determined by flow cytometry. (C, upper plots) The percentages of B cells (CD19⁺) and CD4^-CD8^- T cells (DN T cells, pre-gated on CD3⁺) in the MLN. (C, lower plots) The absolute cell numbers and percentages of DN T cells in the spleen. (D) pDCs (defined as CD11c⁺CD11b^-B220⁺Siglec-H⁺) in the BM and MLN. (E) The absolute numbers and percentages of activated pDCs (MHC-II⁺CD40⁺ or MHC-II^high pDCs) in the spleen. Representative flow cytometry plots of MRL and MRL/lpr are shown. (F and G) The absolute cell numbers and percentages of CD11b^- cDCs (CD11c⁺CD11b^-B220^-Siglec-H^-MHC-II⁺) and CD11b⁺ cDCs (CD11c⁺CD11b⁺B220^-MHC-II⁺) in the spleen and MLN. (H) Percentages of Ly6C⁺ cells in CD11b⁺ cDCs in the spleen and MLN. (I) The percentage of CD40⁺ cells in CD11b⁺ cDCs in the spleen and the percentage of MHC-II^high cells in Ly6C⁺CD11b⁺ cDCs in the MLN. Representative flow cytometry plots of MRL and MRL/lpr are shown. ns: not significant, * P<0.05, ** P<0.01, *** P<0.001, student’s t-test. Data are shown as mean + standard error of the mean (SEM), n = 3 mice in each group.

Fig 2. tRA-induced pathology in the brain and lung.

Starting from 6 weeks of age, MRL and MRL/lpr mice were given, orally and daily, vehicle (canola oil), tRA (6 mg/kg BW), or VARA (6 mg retinol and 0.6 mg tRA per kg BW) till 14 weeks old when tissues were collected. n = 6 mice in each group. (A) H&E stains of the brain. Representative micrographs are shown. Bar equals 100 m. (B) H&E stains of the lung. Representative micrographs are shown. Bar equals 50 μm. Arrows indicate areas of infiltration. (C-D) Leukocyte infiltration scores of the brain (C) and lung (D) according to H&E stains. * P<0.05, ** P<0.01, *** P<0.001, one-way ANOVA. Data are shown as mean + SEM.

Fig 3. tRA-mediated modulation of kidney pathology.

(A-C) Leukocyte infiltration of the tubulointerstitial region. (A) Cell infiltration (CI) scores according to H&E stains. (B) Representative micrographs of H&E stains of the tubulointerstitial region. Bar equals 100 μm. Areas of infiltration are indicated by arrows. (C) Immunohistochemical stains of T cells (CD3-blue), dendritic cells (CD11c-red), and plasma cells (CD138-green). Representative images are shown. Bar equals 100 μm. (D-H) Glomerular analysis (GA). (D) Average GA scores of hypercellularity, mesangial matrix expansion, necrosis, percentage of sclerotic glomeruli, and glomerular crescents. (E) Analysis of proteinuria. The level of total protein in the urine was measured weekly with Chemstrip 2GP. The data of 6- to 8-week-old time points were combined as the early stage, and those of 12- to 14-week-old time points were combined as the late stage. (F) Representative H&E stains showing kidney glomeruli. Bar equals 60 μm. (G) PAS stains showing kidney glomeruli. Bar equals 40 μm. (H) Immunohistochemical stains of IgG (green) and complement C3 (red) deposition in the kidney cortex. Bar equals 200 μm. ns: not significant, * P<0.05, ** P<0.01, *** P<0.001, one-way ANOVA. Data are shown as mean + SEM (n = 6 mice in each group).

Fig 4. tRA-mediated decrease of lymphocyte accumulation in the MLN.

(A) MLN weight, MLN/body weight ratio, and total number of cells in MLN. (B) Absolute numbers of T and B cells in the MLN. * P<0.05, ** P<0.01, *** P<0.001, one-way ANOVA. Data are shown as mean + SEM (n = 6 mice in each group).

Fig 5. Vitamin A-mediated increase of B-cell responses.

(A) Anti-dsDNA IgG, total IgG, and their ratios in the plasma of 8-, 10-, 12- and 14-week-old mice as determined by respective ELISA. One-way ANOVA at each time point was performed but only comparisons between a vitamin A group (either tRA or VARA) and the MRL/lpr vehicle group are labeled in the graphs for simplicity. Although not labeled here, the MRL vehicle group is statistically different from all other groups. Data are shown as mean ± SEM (n = 6). (B) The absolute cell numbers of plasma cells (CD19^-CD27^-CD138⁺CD44⁺) and plasmablasts (CD19^+/lowCD27^+/lowCD138⁺CD44⁺) in the spleen, MLN, and BM at 14 weeks old as measured by flow cytometry. (C) mRNA levels of IL-6, IL-21, and IFNα in the spleen at 14 weeks old as determined by RT-qPCR. Relative quantities (RQ) of cytokine mRNA were normalized to that of L32. The average RQ value of MRL vehicle group was defined as 1. ns: not significant, * P<0.05, ** P<0.01, *** P<0.001, one-way ANOVA; +: P<0.05, student’s t-test. Except for (A), data are shown as mean + SEM (n = 6 mice in each group).

Fig 6. Correlation analysis between blood autoantibody levels and pathological scores.

The area under the curve (AUC) was calculated for the level of anti-dsDNA IgG in the circulation and plotted against pathological scores of the brain, lung and kidney. Spearman correlation tests were performed.

Fig 7. Minimal induction of inflammation with tRA in the absence of an immunogenic environment.

MRL mice were given, orally and daily, vehicle (canola oil) or tRA (6 mg/kg BW) from 6 to 14 weeks of age when tissues were collected. n = 3 mice in each group. (A) H&E stains of the lung. Representative micrographs are shown. Bar equals 400 μm. Areas of infiltration are indicated by arrows. (B) Representative micrographs of H&E stains of the kidney. Bar equals 400 m. (C) Immunohistochemical stains of T cells (CD3-green), dendritic cells (CD11c-red), and plasma cells (CD138-blue). Representative images are shown. Bar equals 200 μm. (D-E) Leukocyte infiltration scores of the lung (D) and kidney (E) according to H&E stains. (F) Analysis of proteinuria. The level of total protein in the urine of 14-week-old mice was measured with Chemstrip 2GP. (G) Anti-dsDNA IgG, total IgG, and their ratios in the plasma of 10-, 12- and 14-week-old mice as determined by respective ELISA. ns: not significant, student’s t-test. Data are shown as mean + SEM or mean ± SEM.