The work reported here uncovers a key molecular mechanism for the involvement of K8/K18 IFs as modulators of hepatic cell mechanics. By comparing shK8b versus H4ev cell adaptation to ECM-derived mechanical cues, using FN-gels of controlled rigidity and measuring cell stiffness with an optical tweezers, we find that a K8/K18 IF loss perturbs the mechanotransduction by: altering the cell mechanosensing, along with a marked actin fiber differential re-organization; interfering with the proper match between cell stiffness and ECM rigidity; and impairing the Rho-ROCK pathway controlling cell stiffness and actin dynamics. Overall, the results reveal a K8/K18 IF contribution to the cell stiffness-ECM mechanosensing interplay through a modulation of Rho-dependent actin cytoskeleton dynamics in simple epithelial cells.
Substratum rigidity has a fundamental impact on molecular events triggered at FAs/integrins, which in turn modulate cell spreading and actin cytoskeleton organization 
. Notably, the range of rigidity where this process can be observed is cell-type dependent. For instance, fibroblasts achieve a maximal spreading and display prominent actin fibers on a substratum of 10 kPa 
, while chondrocytes still maintain a round shape at this rigidity 
. In contrast, mammary epithelial cells undergo a round to spread shape transition between 0.4 and 5 kPa, while actin fibers are detected only above 5 kPa 
. In comparison, MDCK cells achieve a spread shape at rigidity of 1 kPa 
. As expected, the present results indicate that the round to spread shape transition in H4ev cells occurs at rigidity level within the range for typical simple epithelial cells 
. In addition, previous work using fibroblasts and endothelial cells has indicated that the influence of substratum rigidity on cell spreading (shape) and cytoskeletal organization can be modulated by cell-cell contacts, particularly in the case of endothelial cells 
. Conversely, other line of work using the latter cell type has revealed that the rigidity of a substratum adhering on FAs/integrins can affect cell-cell (e.g. actin-associated) junction stability, through a modulation of the Rho-mediated actin cytoskeleton contractility 
. The cells of interest here, derived from simple epithelium, contain not only actin-linked adherens junctions but also keratin-associated desmosomal junctions, which allow formation of a continuous K8/K18 IF network across the cell monolayer. In this context, one can assume that any change in hepatic cell shape (spreading), as result of the K8/K18 IF loss in shK8b cells, is possibly due to an effect on both substratum rigidity and cell-cell contacts. However, the present results show that shK8b cells fail to initiate cell spreading at the lowest rigidity, regardless of cell-cell contacts, where H4ev cells are able to do so, indicating a predominant K8/K18 IFs involvement at events triggered at FAs/integrins rather that at cell-cell contacts at this rigidity level. Moreover, the shape transition in shK8b cells takes place at a higher rigidity threshold, thus providing the first indication that K8/K18 IFs behave as modulators of the hepatic cell mechanosensing of substratum rigidity. In addition, it appears that K8/K18 IFs constitute a modulator of actin fiber organization over a wide range of ECM rigidity, suggesting that the K8/K18 IF-dependent mechanosensing at FAs/integrins is primarily mediated through actin cytoskeleton. In this context, the likely involvement of cell lineage specific-IFs as key regulators in the setting of the actin cytoskeleton-dependent mechanosensing range in simple epithelial cells constitutes an appealing issue.
Stiffness measurements on cells seeded on ECM substrata of various rigidities have been performed using different force detecting systems, including magnetic twisting cytometry, atomic force microscopy and optical tweezers 
. Of particular note, both fibroblasts and mesenchymal stem cells have been found to possess a stiffening behavior that depends on substratum rigidity 
. In line with these findings, our present data point to a stiffening behavior in both H4ev and shK8b epithelial cells, with the difference that hepatic cells lacking K8/K18 IFs stiffen on higher gel rigidity, in correlation with the increased rigidity required to form a cell monolayer. Of note, measurements of alveolar epithelial cell stiffness did not reveal a ECM rigidity-dependent sensitivity 
; however, these cell measurements have been made at ECM rigidity levels well above the normal lung tissue rigidity 
. Actually, there is a saturation point for cell stiffness above a certain level of ECM rigidity 
, stressing the importance of addressing the relationship between ECM rigidity and cell mechanics within physiologically relevant experimental settings. Moreover, the differential actin fiber organization we observe here in hepatic epithelial cells seeded on glass substratum does not correlate directly with the stiffness measurements. Furthermore, both bead size and ECM-coating density have been shown to influence bead adhesion and the associated integrin-mediated mechanical coupling 
. In this context, the stiffness differences observed here in shK8b cells versus H4ev cells upon variation of the bead FN-coating density are likely due to an increased integrin-mediated bead-actin cytoskeleton coupling. Overall, we conclude that the K8/K18 IF loss perturbs the shK8b cell ability to respond to changes in both substratum rigidity and ECM-bead coating density characterized by an altered ability to tune their own stiffness.
There is accumulating evidence indicating that substratum rigidity influences cell stiffness through actomyosin mediation 
. Notably, the control of actomyosin contractility and actin fiber formation is known to be dependent on the Rho-ROCK signaling pathway, which in turn activates different effectors of actin organization 
. Although RhoA is the most characterized among the Rho family members, both RhoA and RhoC can induce formation of fibrillar actin through ROCK activation 
. In addition, the activation of Rho proteins is compartmentalized, meaning that different pools of the same Rho protein can be associated with distinctive actin organization traits 
. The present results revealing a differential actin fiber distribution at the dorsal surface of H4ev versus shK8b cells, and a peripheral localization of actin fibers at the ventral surface of shK8b cells following RhoA overexpression, are consistent with such a Rho compartmentalization. Still, the persistence of some fibrillar actin at the dorsal surface in shK8b cells following ROCK inhibition suggests that alternate signaling pathway(s) can contribute upon a RhoA impairment, in line with previous data revealing a regional fibrillar actin formation through myosin light chain kinase 
. Nevertheless, by combining these findings with those on cell stiffness, it remains that K8/K18 IFs are involved in actin cytoskeleton organization and hepatic cell mechanics, through a strategic modulation of RhoA-ROCK activation.
The present results reveal an altered bead-cytoskeleton coupling and mechanosensing triggered by substratum rigidity at FAs/integrins. Notably, integrin engagement on their respective substratum can regulate Rho activity through the involvement of Src family kinases. Indeed, Src controls RhoA activity by regulating a number of guanine-exchange factors (GEF) and GTPase-activating proteins (GAP), including p190RhoGAP and p190RhoGEF 
. During initial cell spreading, Src enhances p190RhoGAP phosphorylation, leading to a down-regulation of RhoA activity 
, whereas at a later spreading stage, Src can increase RhoA activity through p190RhoGEF and promote actin fiber formation 
. Of additional note, Src activation has been linked to alterations of local cell stiffness at FAs, as measured by FN- or vitronectin-coated bead pulling 
. On this ground, a K8/K18 IF modulation of the Src-mediated bead-mechanical coupling at FAs constitutes an appealing perspective.
We have shown a few years ago that much of the K8/K18 IFs and fibrillar actin are distributed beneath the surface membrane in hepatic cells 
; we have also found a similar distribution for plectin, a cytolinker for IF proteins, actin and microtubule-associated proteins 
. More recently, we have reported that K8/K18 IFs in hepatic cells modulate the sub-cellular localization/interaction of plectin, PKC (“protein kinase C”) and RACK1 (“Receptor of activated C kinase”), in link with the integrin-dependent adhesion and migration 
. In the work reported here, using the same cell model, it appears that K8/K18 IFs are modulators of the RhoA localization-dependent activity. Since RACK1 also comprises a binding site for Src 
and since Src is a major regulator of Rho activation 
, we propose that the interplay between K8/K18 IFs and Rho-mediated actin dynamics occurs through a plectin-RACK1-Src connection at FAs in hepatic cells.
In addition, considering the relation between K8 point mutations and the predisposition to liver cirrhosis 
, and the link between cirrhosis and hepatocarcinogenesis 
, hepatic cells offer a model of choice to address the role of K8/K18 IFs as modulators of the mechanotransduction taking place during the cell tumorigenic process in simple epithelia. Finally, since K8 and K18 constitute the first cytoplasmic IF genes expressed in the embryo 
and since mechanical force plays a key role in governing stem cell differentiation 
, the possibility that K8/K18 IFs intervene in the mechanical stimulation of stem cell lineage emergence during early development constitutes an attractive perspective.