It's strange, but there are mechanisms whereby this type of thing can happen. Basically you have a rate determining step that involves an intermediate species, and that rate determining step must have the usual positive temperature dependence. But the preceding equilibrium that produces the intermediate species could have a negative temperature dependence that dominates.
Let's translate this into concrete terms for the reaction given. A quite plausible mechanism would look like this:
$\ce{NO + O2 <=> ONOO}$ Fast, reversible
$\ce{ONOO + NO -> 2NO2}$ Rate determining step
The equilibrium in the first step is given by
$K_\mathrm{p1} =\dfrac{p^\mathrm{eq}(\ce{ONOO})}{p^\mathrm{eq}(\ce{NO})p^\mathrm{eq}(\ce{O2})}$
Assuming $\ce{ONOO}$ is a minor species, it's formation hardly affects the equilibrium partial pressures of the other species at any time, so we may just use $p(\ce{NO})$ as the nominal nitric oxide partial pressure and similarly for the molecular oxygen. Then we can solve for the partial pressure of the $\ce{ONOO}$ intermediate:
${p^\mathrm{eq}(\ce{ONOO})}=K_\mathrm{p1}(p(\ce{NO}))(p(\ce{O2})$
Now put this result into the rate determining step with rate
$r_2=k_2{p^\mathrm{eq}(\ce{ONOO})p^\mathrm{eq}(\ce{NO})}=k_2K_\mathrm{p1}(p(\ce{NO})^2)(p(\ce{O2})$
Now for the eye-opener: put in the temperature dependence of the equilibrium and rate constants
$K_\mathrm{p1}=K_\mathrm{p1}^0\exp(-\Delta H_1/RT)$
$k_2=k_2^0\exp(-E_{a2}/RT)$
So then
$r_2=k_2^0K_\mathrm{p1}^0\exp(-\dfrac{\color{blue}{E_{a2}+\Delta H_1}}{RT})p(\ce{NO})^2p(\ce{O2})$
The activation energy then contains two terms, one from the inherent rate in Step 2 and the other from the equilibrium in Step 1. The latter is negative if Step 1 is exothermic, as it might well be if a bond is being (temporarily) formed between the reactants. Then if this enthalpy term is the absolutely larger part of the sum ... we see a reversed temperature dependence!