The Native Americans and smallpox
First I want to note that some of the posted answers are not quite accurate. For example, smallpox ravaged the Native American population because they did not have high-affinity MHC molecules for it, which is an evolved trait. MHC molecules must be capable of binding to a large class of peptides, at least weakly, to allow their presentation. There are only a finite number of MHC alleles in a given individual, and they tend to represent the most commonly found classes of foreign organisms.
The Native Americans were equally capable of creating anti-smallpox antibodies as any European, but that process was significantly slowed down due to not having MHC alleles which promote the presentation of smallpox, allowing the virus to accumulate to lethal levels before a sufficiently mature immune response could be mounted. You can bet that the survivors were the ones with MHC alleles that promoted rapid recognition of the virus!
Somatic hypermutation and affinity maturation
The way antibodies work is, B cells undergo both genetic recombination and an intentionally increased mutation rate called somatic hypermutation. This results in a virtually uniform distribution of new antibodies. These new antibodies are actually pretty bad at binding antigens, and at best, only roughly approximate them. This is one reason the first exposure takes so long to eradicate. When an antibody that weakly matches is found and the number of B cells rise, some of them begin to induce additional mutations to their antibody, a process called affinity maturation. The mutations that result in breaking the antibody end up discarded (those B cells die), whereas those that have an even stronger affinity result in the clone of B cells being positively selected for. This goes on and on until the resulting clone of B cells produce extremely high-affinity antigens.
The result of this is that virtually any possible "door" will have a "key", even if it is poorly fitting. Over time (and subsequent exposures), the key is fine-tuned. The only way a persistent antigen could evade an immune response, short of interfering with the immune system is if the antigen is extremely similar or identical to a self-antigen.
Some pathogens can avoid an immune response in indirect ways, such as frequently changing exposed antigens either by switching genes or through rapid mutation, covering the cell surface with self-antigens or innate, non-protein material, tucking conserved regions that act as epitopes deeply inside a protein, releasing molecules that inhibit the immune response, etc.
Foreign antigens resulting in production of anti-self antigens
An interesting thing to note is that, even if the antigen is similar to self, it will still often be detected. This can be highly problematic because selection against anti-self antibodies occurs when the B cell is first developing, during somatic hypermutation, but not during affinity maturation. A weakly binding antigen that is similar to a self-antigen may then stimulate the production of anti-self antibodies since the antigen is different enough from self that B cells specific to it are not negatively selected, but similar enough to self that the changes in affinity maturation are sufficient to convert it into an anti-self antigen (in other words, an antibody x steps away from targeting self-antigens results in negative selection, but affinity maturation results in y changes to the antibody, where y > x). This is one reason why some viruses can trigger auto-immune diseases.
Answering your questions directly
What if there is no antibody in any cell against an antigen?
If for some reason no antibody were even had the slightest affinity for an antigen, then there would be no problem, and a pathogen would trigger an immune response because of another antigen. If for some insane reason every single antigen on a pathogen (there are many) was not even slightly matched, then no humoral immune will be invoked. There still may be a cellular immune response (e.g. for viruses), and the innate immune system would still be active.
Does the immune system has antibodies against all antigens in the world?
The immune system has antibodies that bind a wide range of antigens, albeit weakly, and with high overlap. As a result, you could say that all possible antigens that are not too similar to self are countered by antibodies with at least some level of affinity. There is only a finite number of possible antigenic peptides, and of those, all that's needed is a weak interaction to trigger a full-blown adaptive immune response.
It's like going to an ophthalmologist. They don't have lenses for all possible refractive indexes, but they do have enough samples that, by trying them out, you can see which ones improve your vision. Going through this process allows gradual improvement, and eventually they are able to proscribe a perfect match.
What happens if an antigen came into our body which has no antibody matching for it?
A single antigen, as opposed to a pathogen with many antigens? Ignoring the cellular immune system, nothing would happen. It would be treated virtually identical to self-antigens (i.e. it would be ignored). Depending on how the pathogen behaves, it could be wiped out by the rest of our immune system, contained in a local cyst, or even result in a lethal infection.
