ROS, despite being essential for biological systems  have the potential to cause extensive oxidative damage to cells and tissues if their levels become excessive [33, 34]. At the vascular level ROS can cause oxidative damage of endothelial cells  including DNA strand breakage and inflammation . In addition to ROS, nicotine can equally elicit oxidative stress and tissue injury [35, 36] and has been shown to exacerbate brain edema following focal ischemia [37, 38]. Oxidants in the gaseous phase of cigarette smoke, including nicotine and various ROS species, ([15, 20] can pass through the lung alveolar wall and raise systemic oxidative stress . This can lead to oxidative damage to cells and tissues, including the brain vascular system and the BBB, over a period of sustained exposure to TS (e.g., chronic smokers) and facilitate the pathogenesis and progression of neurological disorders [40–42]. Thus, existing evidence strongly suggests a role for TS-dependent oxidative and inflammatory stress in the development of CNS pathologies. In fact, the cerebrovascular endothelium is highly vulnerable to oxidative stress resulting in loss of BBB function and integrity via altered expression and distribution of intercellular TJ complexes [43, 44].
In this study we assessed and compared the effects of various tobacco products on human BBB endothelial cells in relation to their corresponding oxidative potential. Specifically, several studies have demonstrated that cigarette smoke contains high concentrations of NO which may directly affect the integrity of the BBB. For this purpose we measured ROS as well as NO3−/NO2− content (Figures 2 & 3) of tobacco smoke from 1R5F (ultralight), 3R4F (full flavor), NF (tobacco free) and ultralow nicotine cigarettes. The NO3−/NO2− analysis revealed a direct correlation with the content of tar of the respective cigarettes. However, this did not hold true for NF products, whose tar content (comparable to medium strength cigarettes) produced the least amount of nitrate and nitrite. When we compared the NO3−/NO2− output with corresponding nicotine content, a significant correlation was not found, unless ultralow nicotine brand were removed from the pool (Figure 2C - insets). Together these results suggest that NO3−/NO2− is relatively independent of nicotine content while holding a strong correlation with that of tar.
Tar being a byproduct derived from combustion of tobacco or analogous products, alteration of tobacco (e.g., ultralow nicotine products) or replacement with alternative products (NF cigarette) to reduce nicotine content in a bid to decrease addiction potential, may result in an unwanted increase of nitrate/nitrite output, and risk for health hazard. In fact, tobacco nitrate levels have been previously reported to correlate with the formation of non-specific volatile nitrosamines (e.g., N-nitrosodimethylamine, N-nitroso-diethylamine, N-nitrosoethylmethyl- amine, etc.), and non-volatile Tobacco-Specific Nitrosamines (TSNAs) such as 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK, nicotine-derived nitrosamine ketone) which have been associated with carcinogenicity of tobacco smoke [30, 31].
Interestingly, H2O2 content measured in ultralow nicotine and tobacco free (NF) cigarettes, considered “reduced-exposure” products, was significantly higher than any other brand including medium and full flavor (see Figure 3). Regression analysis of H2O2 also revealed a strong correlation with the tar content of the respective cigarettes but not with that of nicotine unless both products were to be removed from the pool. These results strongly correlate with the increased oxidative stress generated in BBB endothelial cultures (see Figure 3A2) and revealing that the highest level of oxidation is in endothelial cells that are exposed to ultralow and NF cigarette smoke extracts. Interestingly, the oxidative stress potential of 3R4F cigarettes was comparable to that of ultralow and NF cigarettes despite releasing lower amounts of H2O2. This can be attributed to the higher content of nicotine in 3R4F cigarettes since nicotine equally contributes to oxidative stress (Das et al. 2012). Taken together, these results suggest that alteration and/or substitution of tobacco with alternative products in order to reduce nicotine content was responsible for the increased H2O2 output measured in these “denicotinized” cigarette products.
Previous reports by Hossain and co-workers  have shown a dose dependent loss of BBB integrity directly correlating to TS-derived oxidative stress. Furthermore, loss of BBB function and integrity caused by TS exposure was prevented or at least reduced by antioxidant vitamins. These findings by others clearly support our results which outlined a strong correlation between the impairment of tight junction protein expression/distribution and BBB integrity with the oxidative stress generated by the TS extracts. As clearly shown in the results (see Figure 4) BBB endothelial ZO-1 expression and distribution is completely deregulated upon exposure to TS extract from 3R4F, NF and Ultralow Nicotine cigarettes. This is also reflected in the increased BBB permeability to dextran paracellular markers observed under the same conditions.
ZO-1 is a cytoplasmic accessory protein which plays a crucial role in BBB integrity by connecting transmembrane proteins (such as occludin, claudins and JAM) to cytoskeletal proteins and is actively involved in signal transduction and transcriptional modulation [45, 46]. Interestingly, the effect of CSE on ZO-1 expression/distribution reflects the overall oxidative potential of the corresponding cigarettes (see Figure 3), thus suggesting a correlation between TS-dependent oxidative potential and dysregulation of TJs and BBB integrity. ZO-1 TJ protein closely associates with the actin cytoskeletal network. When we observed the actin structure with respect to ZO-1, it appeared intact. In addition, membrane expression of occludin was significantly down-regulated as evidenced by the WB analysis of the corresponding membrane fractions. Similar to ZO-1, membrane distribution of occludin was also altered deteriorating from a homogenous pattern at cell-cell junctions in controls to a patchy distribution in cultures exposed to 3RF4, NF and ultralow nicotine cigarettes. This can be a reflection of the parallel loss of ZO-1 which provides a positioning system and anchoring scaffold for the transmembrane TJ proteins.
In contrast to ZO-1 and occludin, the expression of VE-cadherin and claudin-5 was proportionally increased with respect to the oxidative potential of the corresponding CSE treatment. In fact, as shown in Figure 5, VE-cadherin membrane expression was progressively up-regulated by exposure to 3RF4, NF and ultralow nicotine cigarettes, although statistical significance was proven only for the last two cigarette products. In parallel, claudin-5 membrane expression was similarly up-regulated (see Figure 5B). This is in agreement with emerging evidences suggesting that VE-cadherin controls claudin-5 expression by preventing the nuclear accumulation of FoxO1 and beta-catenin which repress the claudin-5 promoter  thus reducing its expression. Although, these results were surprising, they actually seem to be in agreement with the above mentioned observations. In fact, recent in vitro studies have shown a direct positive correlation between VE-cadherin expression and oxidative stress  suggesting this being part of a cytoprotective response mechanism. In fact, VE-cadherin acts as a master regulator of various endothelial functions including modulation of cell-cell adhesion, angiogenesis, and vascular permeability to leukocytes in response to VCAM-1 activation , whose expression level was also increased (see Figure 6). Note also that an up-regulation of claudin-5 (in this case mediated by VE-cadherin) does not necessarily translate into an improved BBB integrity. Although this is true from a biological standpoint under normal circumstances we have to take into consideration that the mere expression of TJ proteins is not sufficient as a standalone determinant for BBB integrity. Other important factors play a significant role here such as the link between TJ proteins with the cytoskeleton. An important interaction mediated by first order regulatory proteins such as ZO-1 is of critical importance for the positioning and interaction of TJ proteins with their homologues on adjacent endothelial cells. Moreover, although claudin-5 was up-regulated (see Figure 5) the pattern of expression presented as an homogenous distribution throughout the cells and lacked a demarcated membrane localization which does not suggest improvements of cell-cell adhesion. This hypothesis well copes with the evident loss of barrier functions outlined by the increased permeability to dextran markers.
In addition, a similar increase in PECAM-1 expression was observed as well as an increased endothelial release of IL-6 and MMP-2 (see Figure 6). Regarding MMP-2, previous reports by others have shown how ROS regulate the activity of vascular matrix metalloproteinases in vitro including MMP-2 and MMP-9  which have an implication in atherosclerotic plaque stability. Expression and activation of MMP-2 has been demonstrated as a key event in oxidative stress injury to heart  and hyperglycaemia promoted BBB dysfunction . Together these results strengthen the link between tobacco smoke, it’s corresponding oxidative and inflammatory stress, and potential risk for BBB dysfunction. Although outside the scope of the present work, more studies will be necessary to dissect the molecular mechanisms involved in the generation of cellular oxidative stress at the brain microvascular endothelium by CSE and its impact on BBB function and integrity.