PITTSBURGH — The first two defense witnesses in a University of Pittsburgh medical researcher’s homicide trial on Monday targeted the prosecution’s central claim the researcher purposely poisoned his neurologist wife with cyanide last year.

The director of a lab that tested the blood of 41-year-old Dr. Autumn Klein and a California expert on the care of poison victims both testified that they can’t say with certainty that Klein was poisoned.

Allegheny County Assistant District Attorney Lisa Pellegrini contends Dr. Robert Ferrante, 66, killed Klein by lacing her creatine energy drink with cyanide in April 2013. Klein collapsed almost immediately and died three days later.

Pellegrini rested her case after testimony by Dr. Christopher Holstege of the University of Virginia, author of the book, “Criminal Poisoning: Clinical and Forensic Perspectives.”

Holstege testified that Klein’s symptoms ruled out anything but cyanide poisoning. He based his opinion on documents ranging from police reports, records of medical treatments that Klein received after the 911 call until she died three days later, and, finally, on a lab test that found lethal levels of cyanide in blood drawn from her by physicians trying to save her life. Those results weren’t until several days after she died.

The test that found the lethal level was conducted by Quest Diagnostics of Chantilly, Va.

The first defense expert, Dr. Robert Middleberg, is the lab director and vice president of quality assurance at another private lab, NMS Labs of Willow Grove, Pa.

NMS is being paid $5,000 to testify for the defense. But Middleberg — and defense attorney William Difenderfer — noted NMS’ involvement began when the county medical examiner’s office also sent Klein’s blood and plasma samples to also determine the level of cyanide in her blood after a county crime lab test also confirmed cyanide in her blood.

But the NMS test couldn’t be completed because a piece of equipment needed to examine the precise level of the poison in Klein’s blood wasn’t working and because another kind of test on a known amount of cyanide — known as a “control” — came back negative, indicating that the test being used wasn’t reliable.