I am familiar with the situation in Russia and in the US, and your country may have different practices.
Your biggest impediment is that agility of one's brain, and one's ability to learn, really depends on age: it is easier to learn stuff when you are young; that's just biology of the brain. You know this, and other people know this. So while there may be no legal limitations on age, and most civilized countries would explicitly forbid age discrimination, the admission committees in most grad schools would likely screen you out and prefer younger candidates. Remember: the fame of the program depends on both what the faculty publish and on where the graduates are placed, and if they perceive that you will not be placed well five or seven years down the road when you are fifty (and everybody's thinking on the job market would be, "I have plenty of 50-year old professors with three dozen publications and 20 years of teaching experience in my department, why should we bother with a 50-year old Ph.D. graduate who will retire in 10 years?"), they may not take you upfront.
You have not defined what you mean by "professional mathematician"; there's a spectrum from teaching math in high school to working full time on math research at Courant Institute. You would probably be able to get up to Master's if you can pay for your education out of your pocket. But even getting admitted into a Ph.D. would probably require that the department knows you in and out, and knows your dedication to learning math -- which means you will be limited to the school where you can get your Bachelor's and/or Master's.
As a side note, in Russia, there is usually little way for you to get into a Ph.D. (or rather the 3-year research-only program called "candidate of science" in Prussian manner) unless it is right after your first college degree, especially in mathematics where agility of the brain is the primary issue. 90% of the post-graduate students would go to the candidate program back-to-back after their basic degrees (which were 5-6 year programs called "specialist", which are easily mapped to the Western Master's degree), and the remaining 10% were one or two years out of college, and would have done some work in a related area. For that candidate program, you'd have to take formal entrance exams in the subject (not unlike subject GRE), in English, and some sort of philosophy of science (unlike the American process based on application, standardized scores and recommendation letters). You will likely fail the subject exam unless you just had an intense math training in the immediately preceding years, and retained most of the math you've just learned.