Rab6a/a' are important Golgi regulators of pro-inflammatory TNF secretion in macrophages

Abstract

Lipopolysaccharide (LPS)-activated macrophages secrete pro-inflammatory cytokines, including tumor necrosis factor (TNF) to elicit innate immune responses. Secretion of these cytokines is also a major contributing factor in chronic inflammatory disease. In previous studies we have begun to elucidate the pathways and molecules that mediate the intracellular trafficking and secretion of TNF. Rab6a and Rab6a' (collectively Rab6) are trans-Golgi-localized GTPases known for roles in maintaining Golgi structure and Golgi-associated trafficking. We found that induction of TNF secretion by LPS promoted the selective increase of Rab6 expression. Depletion of Rab6 (via siRNA and shRNA) resulted in reorganization of the Golgi ribbon into more compact structures that at the resolution of electron microcopy consisted of elongated Golgi stacks that likely arose from fusion of smaller Golgi elements. Concomitantly, the delivery of TNF to the cell surface and subsequent release into the media was reduced. Dominant negative mutants of Rab6 had similar effects in disrupting TNF secretion. In live cells, Rab6-GFP were localized on trans-Golgi network (TGN)-derived tubular carriers demarked by the golgin p230. Rab6 depletion and inactive mutants altered carrier egress and partially reduced p230 membrane association. Our results show that Rab6 acts on TNF trafficking at the level of TGN exit in tubular carriers and our findings suggest Rab6 may stabilize p230 on the tubules to facilitate TNF transport. Both Rab6 isoforms are needed in macrophages for Golgi stack organization and for the efficient post-Golgi transport of TNF. This work provides new insights into Rab6 function and into the role of the Golgi complex in cytokine secretion in inflammatory macrophages.

Figure 1. Rab6 localizes to Golgi membranes of RAW 264.7 macrophage cells.

(A) After 2 h of LPS incubation significantly increased Rab6 protein level, while not affected clathrin, supporting a role for Rab6 in orchestrating TNF delivery form the TGN in a clathrin-independent manner. Rab6a–GFP localizes mainly on Golgi membranes (B) and its distribution is not perturbed by LPS treatment (C). LPS-activated cells clearly show a TNF surface staining (C, arrowheads). Original optical magnification 63X. Bar: 20 µm (A, B). * = p<0.05, n.s., not significant (pairwise comparisons).

Figure 2. Expression of a single Rab6 inactive isoform inhibits TNF secretion.

(A) After 2 h of LPS incubation Rab6a(T27N)–GFP transfected RAW 264.7 cells do not show plasma membrane staining for TNF (asterisks), while untransfected cells (arrowheads) clearly show surface staining indicative of TNF secretion. (B) Expression of Rab6a(T27N)–GFP results in accumulation of TNF in the Golgi complex. (C) Quantification of the TNF released in the growing medium per each experimental point has been graphed. Control is untransfected cells. Original optical magnification 63X (A, B). Bar: 15 µm (A, B). *** = p<0.001 (pairwise comparisons).

Figure 3. siRNA Rab6 affects Golgi morphology in RAW 264.7 macrophages.

(A) Western blotting analysis confirmed a partial level of Rab6 depletion consisting of a drop of 40% in siRNA Rab6 macrophages, and (B) we also confirmed previous reports of collapsed, and frequently fragmented, Golgi complex compared with the most classical perinuclear shape distribution in control cells, using anti-GM130 as a Golgi marker. (C) Morphological changes on the Golgi complex in siRNA Rab6 macrophages LPS-activated for 2 h were visualized by EM; in control cells, the Golgi complex is formed by a multiple interconnected stacks (a, white arrowheads) approaching one to each other to form a ribbon. In siRNA Rab6 (b, white arrowheads) the Golgi complex consists of a huge single isolate ribbon, ticker and at least three times longer (white arrowheads), as previously characterized . The addition of LPS did not significantly change the Golgi morphology (c, white arrowheads). Original magnification has been reported in each image (and relative inset), and the reference bars present on large images (vertically, top right). (D) After 2 h, LPS-activated RAW 264.7 cells clearly present internal and plasma membrane (surface) TNF staining (asterisks), while in siRNA Rab6 cells the TNF remains trapped in the cell. Original optical magnification 63X (B, D). Bars: 5 µm (B), 1 µm (C), 10 µm (D). N, nucleus; m, mitochondrion; ER, endoplasmic reticulum.

Figure 4. Stable depletion of Rab6 inhibits TNF release and significantly alters the Golgi morphology.

(A) An shRNA Rab6–mCherry lentiviral construct expressed in RAW 264.7 cells induced a significant decrease in Rab6 expression detected by Western blotting analysis and the differences versus non-infected or infected with shRNA mCherry construct were graphed. (B) Morphological analysis confirmed a collapse of the Golgi complex in longer ribbon in shRab6 cells, frequently isolated from each other. Original magnification are indicated below left. Bars: 2 µm (shRNA mCherry), 1 µm (shRNA Rab6–mCherry). (C) Rab6 depletion caused a dramatic inhibition in TNF secretion after LPS activation in the initial 2 h, and recovered after longer experimental time points. **** = p<0.0001 (pairwise comparisons). (D) Dramatic inhibition of surface TNF expression was clearly evident in shRNA Rab6–mCherry expressing cells. Original optical magnification 63X. Bar: 20 µm.

Figure 5. Rab6 localizes with p230 in the Golgi complex.

RAW 264.7 macrophages were transfected with Rab6a–GFP, and fixed with 4% PFA in PBS for 30 min. We found a good co-localization level of Rab6a–GFP with the endogenous p230; LPS incubation for 2 h induced a significant increase of the co-localization of p230 on Rab6a-positive Golgi membranes. Original optical magnification 63X. Bar: 20 µm.

Figure 6. Cryo-immunogold EM reveals an increased co-localization of Rab6 and p230 in LPS-activated RAW 264.7 macrophages.

Control (A) or LPS-activated (B) Rab6–GFP transfected RAW 264.7 macrophages were prepared for cryo-immunogold EM (see Materials and Methods) and then stained for GFP and p230. Unstimulated macrophages did not reveal Rab6 and p230 frequently associated on Golgi/TGN membranes (Aa), or clusters of p230. On the contrary, p230- (Ba) or p230/Rab6-positive membranes (Bb) are more evident in LPS-stimulated macrophages. (C) p230/Rab6-positive membranes increased more then two-fold when macrophages were LPS-activated (from 19.18% to 49.12%), while single p230 tubules decreased (from 73.29% to 40.35%). Bar: 500 nm (A, B).

Figure 7. Rab6/p230-positive vesicular tubular carriers increase in LPS-activated RAW 264.7 macrophages.

RAW 264.7 macrophages transiently expressing Rab6a–GFP (A, right), p230(GRIP)–mCherry (A, left), or co-transfected with Rab6a–GFP and p230(GRIP)–mCherry in control (B, upper panel) or LPS-activated RAW 264.7 macrophages (B, lower panel), were subjected to time-lapse recording. Singles frames derived from supplementary material Movies S1, S2, S3, and S4, respectively, are shown. (C) Quantification of tubule formation positive for Rab6, p230 or RAB6 and p230. Numbers of vesicles and tubules in single cells positive for Rab6-GFP, p230(GRIP)–mCherry and a combination of both that were formed de novo in control and LPS-activated RAW 264.7 macrophages from the Golgi complex within 300 s. The values represents means ± SD of eight cells observed for those expressing Rab6a–GFP (green), p230(GRIP)–mCherry (red), Rab6a–GFP and p230(GRIP)–mCherry (striped), in control or LPS-activated RAW264.7 macrophages. (D) Cells co-transfected with Rab6a(T27N)–GFP and p230(GRIP)–mCherry revealed a decreased level of p230-positive vesicular/tubular carriers. Asterisks highlight examples of tubules arising from the Golgi area. Original optical magnification 63X (A, B, D). Bars: 15 µm. ** = p<0.01 (pairwise comparisons).

Figure 8. siRNA Rab6 affects the p230 localization on the Golgi membranes which is required for TNF secretion.

(A) siRNA Rab6 RAW 264.7 macrophages were co-transfected with SidC_P4C–GFP and GalT–mCherry (a), golgin-97(GRIP)–mCherry (b), p230(GRIP)–mCherry (c). Representative co-localization passing through the line scan (a-c), and plotted on the adjacent graphs (a'-c'), shows a clear decreased localization of p230 on the Golgi membranes (c'), less so efficient for golgin-97 (b'). (B) In the same experimental conditions, RAW 264.7 macrophages were stained for intracellular (a, b) and surface (a''', b''') TNF in the presence of LPS in control (a-a''') and in siRNA Rab6 (b-b''') cells. The addition of TAPI (a''', b''') was used to block TNF cleavage on plasma membrane, otherwise released into the growth medium, and to visualize the TNF staining on the surface of the cells. The depletion of Rab6 inhibits the arrival of TNF on the plasma membrane (b'''), which is concomitant with a partial redistribution of p230 (b'). Original optical magnification 63X. Bar: 10 µm (Aa-b, Ba''', Bb-b''), 15 µm (Bb'''), 20 µm (Ba–a'').

Figure 9. Rab6 and p230 show different BFA-induced cytosolic redistributions.

RAW 264.7 macrophages transftected with Rab6a–GFP were incubated over a time course with BFA 5 µg/ml to follow the cytosolic redistribution of both Rab6 and p230 in control (A) or LPS-activated (B) cells. p230 redistributed faster on LPS-activated macrophages where TNF production had been stimulated, while Rab6a redistribution was not affected by the presence of LPS. These changes, visualized as p230/Rab6 ratios, and relative statistical differences have been plotted (C). Original optical magnification 63X (A, B). Bar: 25 µm (A, B). * = p<0.05, ** = p<0.01 (pairwise comparisons).