The Visual Iconicity Challenge: Evaluating Vision-Language Models on Sign Language Form-Meaning Mapping