Pertussis or whooping cough is a highly contagious respiratory tract disease that is caused by the Gram-negative bacterium Bordetella pertussis. Introduction of whole-cell pertussis (wP) vaccines in the 1940s and 1950s, and later of acellular pertussis (aP) vaccines in the 1980s and 1990s, led to a rapid decline in pertussis incidence. However, as of the 1980s, the numbers of pertussis cases are rising again in many countries, including the Netherlands. Furthermore, whereas pertussis first was a typical disease of young children, it is nowadays predominantly found among adolescents, adults, and, of most concern, immunologically immature small infants. The exact reasons for this re-emergence and altered age distribution are currently not well understood and remain under intense debate. Thus, although pertussis vaccination has been proved successful, the illness remains an important health issue. It is therefore of great importance that novel pertussis vaccines are being developed. The work described in this thesis is mainly focused on the development of an improved wP vaccine. An important problem that should be dealt with is that wP vaccines exhibit considerable reactogenicity. This high reactogenicity is mainly due to the presence of a bio-active molecule, known as lipopolysaccharide (LPS). LPS is one of the major constituents of the Gram-negative bacterial outer membrane and is also known as endotoxin. The endotoxic activity of LPS is mainly determined by the composition of its lipid A moiety. It has been shown that changes in lipid A composition can modulate the endotoxic activity of the LPS. Here, we explored the possibilities of improving wP vaccines by altering their LPS composition. We did this either by expressing LPS-modifying enzymes, by changing LPS biosynthesis in the vaccine strain, or by adding non-toxic LPS derivatives to the vaccine. In addition, we investigated whether the supplementation of aP vaccines with LPS derivatives could be useful for improving these vaccines. We provide proof of principle that altering the LPS composition can indeed be a useful strategy for improving both wP and aP vaccines. We show that modulation of LPS composition cannot only decrease vaccine reactogenicity, but can also increase vaccine efficacy. Furthermore, we show that changes in LPS composition can lead to an altered immune response. Overall, our studies provide several possible strategies that can be used to improve pertussis vaccine.