I doubt SHS would ever cause the effects described, unless the exposure was truly massive. The complaint about marijuana affecting intelligence re: brain development hinges on the central effects of THC itself, while lung cancer from SHS typically involves DNA damage caused by contaminants in the smoke (ranging in size from nitric oxide to benzopyrene). DNA damage is typically considered with a linear-no-threshold-model[1] because any incident of DNA damage can have long-term effects on cancer risk, but neurotransmitters likely do have a threshold for long-term damage, because the rate of neuronal receptor activation has a natural range of variation, and disturbances within that range are unlikely to cause any kind of long-term effect, because they simulate the natural conditions.

1: https://en.wikipedia.org/wiki/Linear_no-threshold_model - I should add that the way we deal with the LNTM in radiation protection is quite similar to the neuron model I just laid out: we look at the natural variation of radiation exposure and don't worry (much) about exposure less than that range.