Apple's "VoiceOver" is their screen reader. It is not a "make a file accessible" (least of all a PDF).
An accessible PDF is a PDF that complies with ISO 14289-1, PDF/UA (the ISO Standard for accessible PDF).
To post process a PDF for accessibility you'd use Acrobat Pro.
To pre-stage much of the work you'd use an appropriate authoring application that supports output of proper tagged PDF.
Some examples of these applications are FrameMaker, InDesign or MS Word.
Tagged PDF is then post-processed with Acrobat Pro to obtain the PDF/UA compliant PDF.
Some useful references/tools:
A copy of ISO 14289-1 ($15 from AIIM), a copy of ISO 32000-1 (Clause ("section") 14 is of note) - information at AIIM website, information at the PDF Association website, the PDF Accessibility Checker (PAC 2), & Callas Software's "pdfGoHTML".
So --- Once you have a PDF/UA compliant PDF ("accessible" PDF) you use Assistive Technology to parse the PDF's structure tree. VoiceOver is a choice; NVDA is perhaps a better choice.
Do remember, no Assistive Technology software defines a file's Accessibility.
Be well... "